Distributed photovoltaic energy management method and system

Abstract

The invention provides an energy management method and system based on distributed photovoltaic, and the method comprises the steps: collecting the real-time operation data of a distributed photovoltaic system, an electrochemical energy storage system, a power grid and a load, and obtaining the standardized real-time data; based on a historical sequence and real-time meteorological information in the standardized real-time data, performing rolling prediction on photovoltaic output and load demands at different time scales in the future by adopting a prediction model; according to the predicted result and the standardized real-time data, generating a charging and discharging power instruction of the energy storage system, a power control instruction of the photovoltaic system and a grid-connected point power plan in the current scheduling period, issuing the instructions to an energy storage converter, a photovoltaic inverter and a grid-connected switch device for execution, and monitoring the actual operation state of each device in the execution process; if the deviation between the actual operation state and the corresponding predicted value exceeds a set threshold value, the instruction generation step is executed again, and self-adaptive energy management integrating coupling prediction, optimization, execution and safety protection is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of photovoltaic energy management, and particularly relates to an energy management method and system based on distributed photovoltaic. BACKGROUND

[0002] The penetration rate of distributed photovoltaic systems in power distribution networks is continuously increasing, and the randomness and intermittency of their output significantly affect the stable operation of power grids. To smooth power fluctuations and improve energy utilization efficiency, electrochemical energy storage systems are often operated in coordination with photovoltaic systems. Existing energy management methods mostly use control strategies based on fixed rules or local optimization, which have the following limitations: first, there is a lack of accurate prediction of photovoltaic output and load demand, resulting in insufficient forward-looking of scheduling plans; second, most methods lack effective closed-loop feedback mechanisms after scheduling instructions are developed, and cannot be adjusted in a timely manner when actual operating conditions deviate from expectations, leading to decreased economic efficiency and stability of system operation; in addition, the protection function of the system is often independent of the main process of energy management, making it difficult to achieve fast and coordinated protection based on the current operating state during power grid failures.

[0003] Therefore, there is an urgent need for an adaptive energy management method that integrates prediction, optimization, execution, and safety protection to achieve efficient, reliable, and intelligent operation of distributed photovoltaic systems. SUMMARY

[0004] The present application provides an energy management method and system based on distributed photovoltaic, to solve the defect that the energy management method in the prior art is single, cannot simultaneously realize the integration of prediction, optimization, execution, and safety protection, and leads to poor management efficiency.

[0005] In a first aspect, the present application provides an energy management method based on distributed photovoltaic, comprising:

[0006] Collecting and standardizing the real-time operating data of the distributed photovoltaic system, the electrochemical energy storage system, the power grid, and the load to obtain standardized real-time data;

[0007] Based on the historical sequence in the standardized real-time data and real-time weather information, using a prediction model to perform rolling prediction of photovoltaic output and load demand at different time scales in the future;

[0008] Constructing a multi-objective optimization function with the optimization objectives of system operation economic efficiency, energy self-consistency rate, and power grid impact minimization, and generating the charge and discharge power instructions of the energy storage system, the power control instructions of the photovoltaic system, and the grid-connected point power plan within the current scheduling period according to the results of the rolling prediction and the standardized real-time data;

[0009] The charge-discharge power instruction, power control instruction and grid-connected point power plan are respectively sent to the energy storage converter, photovoltaic inverter and grid-connected switch device for execution, and the actual operation state of each device is monitored during execution;

[0010] If the deviation between the actual operation state and the corresponding predicted value exceeds the set threshold value, the charge-discharge power instruction of the energy storage system, the power control instruction of the photovoltaic system and the grid-connected point power plan in the current scheduling period are regenerated;

[0011] During the whole process of data acquisition, instruction generation and execution, island detection, harmonic analysis and device state diagnosis are performed in parallel. When the island detection identifies a power grid fault, the grid-connected switch device is controlled to be cut off according to the power control instruction and the grid-connected point power plan, and the energy storage converter is controlled to be switched to an off-grid operation mode according to the charge-discharge power instruction to supply power to the key load.

[0012] According to the energy management method based on distributed photovoltaic provided by the application, the future photovoltaic output and load demand of different time scales are predicted by using a prediction model, which includes:

[0013] A machine learning prediction model is used, and historical photovoltaic output data, historical load data, historical and real-time weather information and date types are used as input features. The weather information at least includes irradiance, ambient temperature and cloud cover;

[0014] Based on the input features, the prediction of the ultra-short time scale is performed for real-time scheduling control, the prediction of the short time scale is performed for generating day-ahead plans, and the prediction of the medium time scale is performed for supporting the long-term operation planning of the system;

[0015] The prediction results of each time scale are updated in a rolling prediction manner, and the latest data is integrated at the beginning of each scheduling period.

[0016] According to the energy management method based on distributed photovoltaic provided by the application, the charge-discharge power instruction of the energy storage system, the power control instruction of the photovoltaic system and the grid-connected point power plan in the current scheduling period are generated according to the results of the rolling prediction and the standardized real-time data, which includes:

[0017] The photovoltaic output and load demand results of the rolling prediction are used as basic input, and the state of charge of the energy storage, the grid price signal and the grid-connected point power limit in the standardized real-time data are combined;

[0018] An optimization model is constructed to minimize the operation cost, maximize the self-generation and self-use rate, and minimize the grid power fluctuation, and the model is solved;

[0019] According to the solving result, the charging and discharging power value of the energy storage system is distributed to form the charging and discharging power instruction; the output power limit value or power factor of the photovoltaic inverter is set to form the power control instruction; and the active and reactive power targets of the grid-connected point are calculated to form the grid-connected point power plan.

[0020] According to the energy management method based on the distributed photovoltaic provided by the application, the optimization model aiming at minimizing the operation cost, maximizing the self-generation and self-use rate and minimizing the power fluctuation of the power grid is constructed, which comprises:

[0021] A target function including the electricity purchase cost, the equipment depreciation cost and the maintenance cost is established;

[0022] The state of charge constraint of the energy storage system, the charging and discharging power constraint, the grid-connected point power constraint and the equipment operation constraint are set;

[0023] The linear programming or mixed integer programming algorithm is used to solve the multi-objective optimization problem to obtain a Pareto optimal solution set;

[0024] According to the real-time power grid price signal and the system operation preference, the optimal scheduling scheme is selected from the solution set.

[0025] According to the energy management method based on the distributed photovoltaic provided by the application, the charging and discharging power instruction of the energy storage system, the power control instruction of the photovoltaic system and the grid-connected point power plan in the current scheduling period are regenerated, which comprises:

[0026] The actual output of the photovoltaic system, the actual demand of the load and the actual state of charge of the energy storage system are monitored in real time, and are compared with the corresponding predicted values in the current scheduling period;

[0027] When the deviation absolute value of any comparison result exceeds the preset threshold value of the corresponding category, it is determined that the system operation state deviates significantly;

[0028] In response to the significant deviation, the optimization process is triggered to be re-executed, and the multi-objective optimization function is re-solved based on the latest actual operation state data and the updated prediction result;

[0029] According to the re-solved result, the updated charging and discharging power instruction of the energy storage system, the power control instruction of the photovoltaic system and the grid-connected point power plan are generated and immediately issued to the corresponding equipment for execution.

[0030] According to the energy management method based on the distributed photovoltaic provided by the application, the grid-connected switch device is controlled to be cut off from the grid connection according to the power control instruction and the grid-connected point power plan, and the energy storage converter is controlled to be switched to the off-grid operation mode according to the charging and discharging power instruction to supply power to the critical load, which comprises:

[0031] After detecting the grid fault signal, the output of the photovoltaic inverter and the energy storage converter is adjusted according to the active power instruction in the grid-connected point power plan, so as to stabilize the grid-connected point voltage and frequency;

[0032] A trip instruction is sent to the grid-connected switch device to disconnect the distributed photovoltaic system from the public grid;

[0033] After confirming that the grid connection has been disconnected, an off-grid operation mode instruction is sent to the energy storage converter, and the energy storage converter is controlled to work in a voltage source mode according to the latest charge and discharge power instruction, so as to establish a stable voltage and frequency independent grid for the key load in the island;

[0034] The output power of the photovoltaic inverter and the discharge power of the energy storage converter are continuously adjusted to match the real-time demand of the key load and maintain the power balance of the off-grid system.

[0035] According to the energy management method based on distributed photovoltaic provided by the application, the collected and standardized data are obtained, including:

[0036] The abnormal values and missing values in the collected data are identified and removed, and the missing values are supplemented by adjacent data or filled with historical same period data;

[0037] The original data with different sampling frequencies and dimensions are converted into time series data with uniform time granularity and standard units;

[0038] The processed data is normalized to fall within a uniform numerical interval, forming the standardized real-time data.

[0039] According to the energy management method based on distributed photovoltaic provided by the application, the energy management method based on distributed photovoltaic further includes:

[0040] The photovoltaic power generation, energy storage charge and discharge, and exchange of electricity with the grid are recorded to the distributed ledger;

[0041] Based on the smart contract, the electricity fee settlement with the grid and the distributed photovoltaic electricity transaction with the adjacent user are automatically executed;

[0042] In the off-grid operation mode, the power supply priority of the key load is dynamically adjusted based on the blockchain record.

[0043] According to the energy management method based on distributed photovoltaic provided by the application, the energy management method based on distributed photovoltaic further includes:

[0044] The key performance indicators of the system operation are calculated periodically, including the energy self-consistency rate, the electricity cost saving rate, and the grid power fluctuation rate;

[0045] Based on the historical operation data and performance evaluation results, the parameters of the prediction model and the optimization model are automatically tuned;

[0046] The weight distribution of each time scale prediction and the adjustment strategy of power instruction are optimized by a reinforcement learning algorithm.

[0047] In a second aspect, the present application provides an energy management system based on distributed photovoltaic, comprising:

[0048] A collection module is configured to collect and standardize real-time operation data of the distributed photovoltaic system, the electrochemical energy storage system, the power grid and the load to obtain standardized real-time data.

[0049] A prediction module is configured to use a prediction model to perform rolling prediction on photovoltaic output and load demand at different time scales in the future based on historical sequences in the standardized real-time data and real-time meteorological information.

[0050] A generation module is configured to construct a multi-objective optimization function with system operation economy, energy self-consistency rate and power grid impact minimization as optimization objectives, and generate charging and discharging power instructions of the energy storage system, power control instructions of the photovoltaic system and grid-connected point power plans in a current scheduling period according to the rolling prediction results and the standardized real-time data.

[0051] An execution module is configured to send the charging and discharging power instructions, the power control instructions and the grid-connected point power plans to the energy storage converter, the photovoltaic inverter and the grid-connected switch device respectively for execution, and monitor actual operation states of each device during the execution process; if a deviation between the actual operation states and corresponding predicted values exceeds a set threshold, the charging and discharging power instructions of the energy storage system, the power control instructions of the photovoltaic system and the grid-connected point power plans in the current scheduling period are regenerated.

[0052] A monitoring module is configured to perform island detection, harmonic analysis and device state diagnosis in parallel throughout the whole process of data collection, instruction generation and execution; when the island detection identifies a power grid failure, the grid-connected switch device is controlled to be disconnected from the grid according to the power control instructions and the grid-connected point power plans, and the energy storage converter is controlled to be switched to an off-grid operation mode according to the charging and discharging power instructions to supply power to critical loads.

[0053] In a third aspect, the present application further provides an electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the energy management method based on distributed photovoltaic as described in any of the above aspects when executing the program.

[0054] In a fourth aspect, the present application further provides a non-transitory computer readable storage medium having a computer program stored thereon, wherein the computer program is executable by a processor to implement the energy management method based on distributed photovoltaic as described in any of the above aspects.

[0055] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the energy management method based on distributed photovoltaic as described above.

[0056] The energy management method and system based on distributed photovoltaics provided by this invention, through unified and standardized processing of multi-source operating data, and the generation of coordinated power commands based on multi-objective optimization using fused prediction and real-time data, effectively avoids the problems of conflicting control objectives and asynchronous commands in traditional methods, thus improving the overall control accuracy and system stability. By leveraging multi-timescale prediction and online rolling optimization, more accurate scheduling plans can be generated. Simultaneously, deviation monitoring and re-triggering mechanisms are introduced during execution, enabling the system to adjust online in response to emergencies, significantly improving the system's economic efficiency and adaptability. Through parallel safety monitoring and diagnosis throughout the process, and by effectively utilizing existing power commands during protection actions, a smooth and rapid switching between grid-connected and off-grid modes is achieved during fault events. This ensures the reliability of power supply to critical loads while avoiding severe impacts on system operating points during protection processes, improving the overall safety and availability of the system. Attached Figure Description

[0057] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0058] Figure 1 This is a flowchart illustrating the energy management method based on distributed photovoltaics provided in this embodiment;

[0059] Figure 2 This is a schematic diagram of the structure of the energy management system based on distributed photovoltaic provided in this embodiment;

[0060] Figure 3 This is a schematic diagram of the structure of the electronic device provided in this embodiment. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0062] Figure 1This is a flowchart illustrating the energy management method based on distributed photovoltaics provided in this embodiment.

[0063] like Figure 1 As shown, the energy management method based on distributed photovoltaics provided in this embodiment of the invention is applicable to distributed photovoltaic-electrochemical energy storage collaborative operation systems in scenarios such as industrial and commercial parks, large public buildings, and distributed microgrids. The core components of this system include: a distributed photovoltaic array (including photovoltaic module strings and combiner boxes), an electrochemical energy storage system (such as lithium battery packs and supporting battery management systems (BMS)), grid connection and power conversion equipment (grid-connected switches, photovoltaic inverters, and energy storage converters (PCS)), a load cluster (divided into critical loads such as emergency lighting and core production equipment, and general loads such as air conditioning and ordinary lighting), and a data acquisition and communication system (including sensors, edge computing gateways, and industrial communication networks). The method mainly includes the following steps:

[0064] 101. Collect and standardize the real-time operating data of distributed photovoltaic systems, electrochemical energy storage systems, power grids and loads to obtain standardized real-time data.

[0065] Specifically, through distributed deployment of sensing devices and acquisition modules, comprehensive system operation data and environmental data are acquired:

[0066] A DC current / voltage sensor is deployed in the photovoltaic combiner box to collect the series current of each component and the total voltage of the array; an AC power sensor is deployed at the output of the photovoltaic inverter to collect the output active power, reactive power and operating efficiency, with the sampling frequency set to 10Hz; a meteorological monitoring unit is set up near the photovoltaic array to collect real-time irradiance, ambient temperature and cloud cover data, with the sampling frequency set to 1Hz.

[0067] The battery management system (BMS) collects data on battery cluster voltage, individual cell voltage and temperature, state of charge (SOC) of the energy storage system, and charge / discharge current. The energy storage converter (PCS) collects AC side voltage, frequency, and charge / discharge power, with the sampling frequency set to 5Hz.

[0068] Install intelligent monitoring and control terminals at the grid connection point to collect grid voltage, frequency, active / reactive power exchanged with the grid, and real-time electricity price signals (peak-valley-flat indicators), with a sampling frequency of 1Hz; deploy power sensors in the load distribution circuit according to the "critical / general" type to collect the real-time power consumption of each load, with a sampling frequency of 5Hz.

[0069] All collected data is transmitted to the edge computing gateway via industrial communication protocols such as Modbus and MQTT to achieve centralized data aggregation.

[0070] The raw data undergoes a three-stage processing flow: anomaly removal, format normalization, and standardization. Specifically: Statistical methods are used to identify and remove outliers (such as extreme values ​​exceeding normal operating ranges) in data like photovoltaic power and grid voltage; for missing values, high-frequency sampled power data is linearly interpolated using adjacent data, while low-frequency sampled meteorological and electricity price data is filled with historical data from the same period under the same conditions. Data on photovoltaic output, energy storage SOC, and load power at different sampling frequencies are converted into time-series data with a unified time granularity through timestamp alignment; various physical quantities are converted to industry-standard units (e.g., power is standardized to kW, current to A) to eliminate dimensional differences. A maximum-minimum normalization method is used to map the processed data to a unified numerical range, forming standardized real-time data and ensuring consistent data magnitude.

[0071] By collecting data from all dimensions to cover the system's operational status, and by eliminating data noise and format barriers through standardized processing, high-quality input is provided for subsequent prediction and optimization processes, avoiding control deviations caused by data quality issues, and solving the pain points of scattered and inconsistent data.

[0072] 102. Based on historical sequences and real-time meteorological information in standardized real-time data, a predictive model is used to make rolling predictions of photovoltaic power output and load demand at different time scales in the future.

[0073] Specifically, a machine learning prediction model (such as a Long Short-Term Memory network model) is used, with input features fused from historical data and real-time dynamic data. Historical features include historical photovoltaic power output sequences and historical load demand sequences extracted from standardized real-time data; real-time dynamic features include real-time meteorological information (irradiance, ambient temperature, cloud cover) and date type (encoded as weekday / holiday, season). The model is trained and optimized using historical system operating data (at least 6 months) to ensure that the predictions are adapted to the characteristics of the scenario.

[0074] Based on the above model, three types of time-scale forecasts are performed in parallel to match different scheduling needs: Ultra-short-term forecasting: Predicts photovoltaic output and load demand in the next few minutes, with the results used for dynamic adjustment of instructions in real-time scheduling to cope with instantaneous changes in operating conditions. Short-term forecasting: Predicts photovoltaic output and load demand for the next 24 hours (hourly step), with the results used to generate the basic scheduling plan for the next day. Medium-term forecasting: Predicts the average daily photovoltaic output and load demand for the next month (daily step), with the results supporting the system's medium- and long-term operational planning (such as energy storage capacity optimization).

[0075] Using a fixed scheduling cycle as the update node, at the beginning of each cycle, the latest collected standardized real-time data is integrated into the prediction model to regenerate prediction results at each time scale, covering historical prediction values ​​and ensuring that the prediction is synchronized with the actual operating trend.

[0076] By using multi-timescale predictions to cover different scheduling scenarios and combining a rolling update mechanism to achieve dynamic calibration of prediction results, this approach solves the problems of "insufficient foresight in prediction and disconnection from real-time operating conditions" in traditional methods, providing reliable trend guidance for precise scheduling.

[0077] 103. Construct a multi-objective optimization function with the optimization objectives of system operation economy, energy self-sufficiency rate and grid impact minimization, and generate charging and discharging power commands for the energy storage system, power control commands for the photovoltaic system and grid connection point power plans for the current scheduling cycle based on the rolling forecast results and standardized real-time data.

[0078] A multi-objective optimization function is constructed with economy, self-consistency, and safety as its core objectives, while rigid constraints are set:

[0079] (1) Multi-objective function design: Operational economy objective: Minimize the total system cost, including grid purchase cost, depreciation cost of photovoltaic and energy storage equipment, and operation and maintenance cost. Energy self-sufficiency objective: Maximize the self-consumption rate, i.e., the ratio of photovoltaic self-generated power to load demand. Grid security objective: Minimize the power fluctuation amplitude at the grid connection point and reduce the impact on the grid.

[0080] (2) Constraints: Energy storage constraint: Energy storage SOC is between 20% and 80%, and charging and discharging power does not exceed the rated power of the PCS. Photovoltaic constraint: The output power of the photovoltaic inverter does not exceed the maximum output of the array, and the power factor meets the grid requirements. Grid constraint: Active and reactive power at the grid connection point do not exceed the grid-given limits. Equipment constraint: The operating parameters of equipment such as inverters and PCS are within safe thresholds.

[0081] Linear programming or mixed integer programming algorithms are used to solve multi-objective optimization problems and obtain Pareto optimal solution sets. Combining real-time electricity price signals with system operation preferences (such as peak-hour priority economy and high-fault-incidence period priority safety), the optimal scheduling scheme is selected from the solution set.

[0082] Based on the optimal solution, three types of core instructions are generated: charging and discharging power instructions for the energy storage system (specifying the charging / discharging mode and power value), power control instructions for the photovoltaic system (setting the inverter output power limit or power factor), and grid connection point power plan (specifying the active / reactive power target).

[0083] Through multi-objective collaborative optimization and dynamic scheme selection, conflicts in control objectives among photovoltaic, energy storage, and power grid are avoided, and the generated instructions have coordination consistency and global optimality.

[0084] 104. The charging and discharging power command, power control command and grid connection point power plan are respectively sent to the energy storage converter, photovoltaic inverter and grid connection switchgear for execution, and the actual operating status of each device is monitored during the execution process.

[0085] An edge gateway-device controller hierarchical distribution mechanism is adopted: the edge gateway categorizes generated instructions according to device type and distributes them to the controllers of energy storage converters, photovoltaic inverters, and grid-connected switching equipment via the industrial communication network, driving the devices to execute operations. During execution, core operating parameters are monitored in real time: actual output of the photovoltaic system, actual load demand, actual SOC of energy storage, and actual power at the grid connection point, etc. The monitoring data is synchronously transmitted back to the edge gateway, realizing closed-loop perception of the instruction execution status.

[0086] 105. If the deviation between the actual operating status and the corresponding predicted value exceeds the set threshold, the charging and discharging power command of the energy storage system, the power control command of the photovoltaic system, and the grid connection point power plan for the current scheduling cycle will be regenerated.

[0087] Specifically, different thresholds are preset according to parameter type (e.g., PV output deviation ±10%, energy storage SOC deviation ±2%, grid connection point power deviation ±5%), and the deviation needs to be sustained for multiple sampling cycles to avoid false triggering by instantaneous disturbances. When any parameter deviation exceeds the threshold, the "data update-re-optimize-instruction replacement" process is immediately triggered: the rolling prediction results are updated based on the latest real-time data, the optimization function is re-solved to generate a correction instruction, marked as "emergency execution" and sent to the equipment to replace the original instruction, while the reason for the deviation is recorded.

[0088] Through a closed-loop mechanism of instruction-execution-monitoring-correction, the system's operating status is dynamically and adaptively adjusted, thereby enhancing the system's ability to cope with sudden disturbances.

[0089] 106. During the entire process of data acquisition, command generation and execution, islanding detection, harmonic analysis and equipment status diagnosis are performed in parallel. When islanding detection identifies a grid fault, the grid-connected switching equipment is controlled to disconnect the grid connection according to the power control command and the grid connection point power plan. The energy storage converter is also controlled to switch to off-grid operation mode according to the charging and discharging power command to supply power to the critical load.

[0090] Specifically, throughout the entire process of data acquisition, command generation, and execution, the safety monitoring module operates synchronously, performing three core tasks: First, it uses methods such as voltage phase mutation and frequency change rate monitoring to identify in real time whether a grid fault has caused islanding. Second, it calculates the total harmonic distortion rate by decomposing the grid connection point current signal to determine whether it meets grid harmonic standards. Third, it monitors equipment parameters such as inverter temperature and energy storage battery voltage, predicts potential faults through trend analysis, and issues tiered early warnings.

[0091] When an islanding detection confirms a grid fault, the system should respond quickly according to the following sequence: Based on the grid connection point power plan, adjust the output of the photovoltaic inverter and energy storage converter to briefly stabilize the grid connection point voltage and frequency; issue a trip command to the grid-connected switching equipment to disconnect the system from the public grid, and confirm the successful disconnection through voltage monitoring; send an off-grid operation command to the energy storage converter to make it operate in voltage source mode according to the charging and discharging power command, and establish an independent power supply grid for critical loads; dynamically adjust the photovoltaic output and energy storage discharge power to match the real-time demand of critical loads and ensure the stable operation of the off-grid system.

[0092] By deeply integrating safety monitoring and energy management processes, a seamless connection is achieved between rapid identification, stable transition, mode switching, and load protection under fault conditions. This solves the problems of large switching impact and power loss of critical loads caused by the disconnect between protection and scheduling in traditional methods, thereby improving system safety and power supply reliability.

[0093] Furthermore, based on the above embodiments, this embodiment also includes: recording photovoltaic power generation, energy storage charging and discharging, and electricity exchanged with the grid to a distributed ledger; automatically executing electricity bill settlement with the grid and distributed photovoltaic power transactions with neighboring users based on smart contracts; and dynamically adjusting the power supply priority of key loads based on blockchain records in off-grid operation mode.

[0094] Specifically, core data such as photovoltaic power generation, energy storage charging and discharging, and electricity exchanged with the grid are recorded in real time to a distributed ledger, and hash encryption is used to ensure that the data is tamper-proof and traceable. Pre-set electricity bill settlement contracts and peer-to-peer transaction contracts are implemented—electricity billing with the grid is automatically completed when the settlement cycle arrives, and electricity transactions are automatically completed with nearby users when photovoltaic output is in surplus. A priority list of critical loads is stored in the blockchain, and the power supply sequence is automatically adjusted according to the list when operating off-grid to ensure core needs are met.

[0095] The model can also be iteratively optimized periodically (e.g., monthly): key indicators such as energy self-sufficiency rate, electricity cost savings rate, and grid power fluctuation rate are calculated, and the deviations between actual and target values ​​are compared. Based on historical data and performance deviations, the network weights of the prediction model are automatically corrected, and the weights of the model's objective function are optimized. Through reinforcement learning algorithms, the weight allocation and deviation adjustment strategies for multi-timescale predictions are optimized to improve the long-term performance of the system.

[0096] Blockchain enables data trustworthiness and transaction automation, reducing settlement disputes and transaction costs; model self-optimization forms a virtuous cycle of operation-evaluation-iteration, ensuring that the system maintains its optimal operating state in the long term.

[0097] Based on the same general inventive concept, this invention also protects an energy management system based on distributed photovoltaics. The energy management system based on distributed photovoltaics described below and the energy management method based on distributed photovoltaics described above can be referred to in correspondence.

[0098] Figure 2 This is a schematic diagram of the structure of the energy management system based on distributed photovoltaic provided by the present invention.

[0099] like Figure 2 As shown in the figure, this embodiment provides an energy management system based on distributed photovoltaics, including:

[0100] The data acquisition module 201 is used to acquire and standardize the real-time operating data of distributed photovoltaic systems, electrochemical energy storage systems, power grids and loads to obtain standardized real-time data.

[0101] Prediction module 202 is used to make rolling predictions of photovoltaic power output and load demand at different time scales in the future based on historical sequences and real-time meteorological information in standardized real-time data and using a prediction model.

[0102] The generation module 203 is used to construct a multi-objective optimization function with the optimization objectives of system operation economy, energy self-sufficiency rate and grid impact minimization, and generate the charging and discharging power command of the energy storage system, the power control command of the photovoltaic system and the grid connection point power plan within the current scheduling cycle based on the rolling forecast results and standardized real-time data.

[0103] The execution module 204 is used to send the charging and discharging power command, power control command and grid connection point power plan to the energy storage converter, photovoltaic inverter and grid connection switchgear respectively for execution, and monitor the actual operating status of each device during the execution process; if the deviation between the actual operating status and the corresponding predicted value exceeds the set threshold, the charging and discharging power command of the energy storage system, the power control command of the photovoltaic system and the grid connection point power plan for the current scheduling cycle are regenerated.

[0104] The monitoring module 205 is used to perform islanding detection, harmonic analysis and equipment status diagnosis in parallel throughout the entire process of data acquisition, command generation and execution. When islanding detection identifies a grid fault, it controls the grid-connected switching equipment to disconnect the grid connection according to the power control command and the grid connection point power plan, and controls the energy storage converter to switch to off-grid operation mode according to the charging and discharging power command to supply power to the critical load.

[0105] Figure 3 This is a schematic diagram of the structure of the electronic device provided in this embodiment.

[0106] like Figure 3As shown, the electronic device may include a processor 301, a communication interface 302, a memory 303, and a communication bus 304. The processor 301, communication interface 302, and memory 303 communicate with each other via the communication bus 304. The processor 301 can call logical instructions stored in the memory 303 to execute an energy management method based on distributed photovoltaics.

[0107] Furthermore, the logical instructions in the aforementioned memory 303 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0108] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the energy management method based on distributed photovoltaic provided by the above methods.

[0109] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the energy management methods based on distributed photovoltaics provided by the methods described above.

[0110] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0111] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0112] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An energy management method based on distributed photovoltaic power, characterized in that, include: Real-time operational data from distributed photovoltaic systems, electrochemical energy storage systems, power grids, and loads are collected and standardized to obtain standardized real-time data. Based on the historical sequences and real-time meteorological information in the standardized real-time data, a prediction model is used to make rolling predictions of photovoltaic power output and load demand at different time scales in the future. A multi-objective optimization function is constructed with the optimization objectives of system operation economy, energy self-sufficiency rate and minimization of grid impact. Based on the rolling forecast results and the standardized real-time data, the charging and discharging power command of the energy storage system, the power control command of the photovoltaic system and the grid connection point power plan are generated for the current scheduling cycle. The charging and discharging power command, power control command, and grid connection point power plan are respectively sent to the energy storage converter, photovoltaic inverter, and grid connection switchgear for execution, and the actual operating status of each device is monitored during the execution process; If the deviation between the actual operating status and the corresponding predicted value exceeds the set threshold, the charging and discharging power command of the energy storage system, the power control command of the photovoltaic system, and the grid connection point power plan for the current scheduling cycle will be regenerated. Throughout the data acquisition, instruction generation, and execution process, islanding detection, harmonic analysis, and equipment status diagnosis are performed in parallel. When the islanding detection identifies a grid fault, the grid-connected switching equipment is controlled to disconnect the grid connection according to the power control instruction and the grid connection point power plan. The energy storage converter is also controlled to switch to off-grid operation mode according to the charging and discharging power instruction to supply power to critical loads.

2. The energy management method based on distributed photovoltaic power according to claim 1, characterized in that, The method of using a predictive model to make rolling forecasts of photovoltaic power output and load demand at different time scales in the future includes: A machine learning prediction model is used, with historical photovoltaic power output data, historical load data, historical and real-time meteorological information, and date type as input features. The meteorological information includes at least irradiance, ambient temperature, and cloud cover. Based on the input features, predictions are performed at ultra-short-term time scales for real-time scheduling and control; predictions are performed at short-term time scales to generate day-ahead plans; and predictions are performed at medium-term time scales to support long-term operational planning in the system. The forecasts are updated at each time scale using a rolling forecasting approach, and the latest data is integrated at the beginning of each scheduling cycle.

3. The energy management method based on distributed photovoltaic power according to claim 1, characterized in that, The process of generating charging and discharging power commands for the energy storage system, power control commands for the photovoltaic system, and grid connection point power plans for the current scheduling cycle based on the results of the rolling forecast and the standardized real-time data includes: The results of the rolling forecast of photovoltaic output and load demand are used as the basic input, and combined with the energy storage state of charge, grid price signal and grid connection point power limit in the standardized real-time data; An optimization model is constructed with the objectives of minimizing operating costs, maximizing self-consumption rate, and minimizing grid power fluctuations, and the model is solved. Based on the solution results, the charging and discharging power values ​​of the energy storage system are allocated to form the charging and discharging power command; the output power limit or power factor of the photovoltaic inverter is set to form the power control command; the active and reactive power targets of the grid connection point are calculated to form the grid connection point power plan.

4. The energy management method based on distributed photovoltaic power according to claim 3, characterized in that, The optimization model, which aims to minimize operating costs, maximize self-consumption rate, and minimize grid power fluctuations, includes: Establish an objective function that includes electricity purchase cost, equipment depreciation cost, and maintenance cost; Set constraints on the state of charge of the energy storage system, charging and discharging power, grid connection point power, and equipment operation; Linear programming or mixed integer programming algorithms are used to solve multi-objective optimization problems to obtain Pareto optimal solution sets; The optimal scheduling scheme is selected from the solution set based on real-time grid electricity price signals and system operation preferences.

5. The energy management method based on distributed photovoltaic power according to claim 1, characterized in that, The process of regenerating the charging and discharging power commands for the energy storage system, the power control commands for the photovoltaic system, and the grid connection point power plan for the current scheduling cycle includes: Real-time monitoring of the actual output of the photovoltaic system, the actual load demand, and the actual state of charge of the energy storage system, and comparison with the corresponding predicted values ​​within the current scheduling cycle; When the absolute value of the deviation of any comparison result exceeds the preset threshold of the corresponding category, it is determined that the system operating status has deviated significantly. In response to the significant deviation, the optimization process is re-executed, and the multi-objective optimization function is re-solved based on the latest actual operating status data and updated prediction results. Based on the results of the re-solution, updated energy storage system charging and discharging power commands, photovoltaic system power control commands, and grid connection point power plans are generated and immediately sent to the corresponding equipment for execution.

6. The energy management method based on distributed photovoltaic power according to claim 1, characterized in that, The step of controlling the grid-connected switching equipment to disconnect the grid connection according to the power control command and the grid connection point power plan, and controlling the energy storage converter to switch to off-grid operation mode according to the charging and discharging power command to supply power to the critical load includes: After detecting a grid fault signal, the output of the photovoltaic inverter and energy storage converter is adjusted according to the active power command in the grid connection point power plan to stabilize the grid connection point voltage and frequency. Send a trip command to the grid-connected switching equipment to disconnect the distributed photovoltaic system from the public power grid; After confirming that the grid connection has been disconnected, an off-grid operation mode command is sent to the energy storage converter, and according to the latest charging and discharging power command, it is controlled to operate in voltage source mode to establish an independent power grid with stable voltage and frequency for the key loads on the island. The output power of the photovoltaic inverter and the discharge power of the energy storage converter are continuously adjusted to match the real-time needs of critical loads and maintain the power balance of the off-grid system.

7. The energy management method based on distributed photovoltaic power according to claim 1, characterized in that, The process of data collection and standardization yields standardized real-time data, including: Identify and remove outliers and missing values ​​from the collected data, and fill missing values ​​with adjacent data or historical data from the same period. Convert raw data with different sampling frequencies and units into time series data with uniform time granularity and standard units; The processed data is normalized to fall within a uniform numerical range, thus forming the standardized real-time data.

8. The energy management method based on distributed photovoltaic power according to claim 1, characterized in that, Also includes: Record photovoltaic power generation, energy storage charging and discharging, and electricity exchanged with the grid to a distributed ledger; Based on smart contracts, automatic execution of electricity bill settlement with the power grid and distributed photovoltaic power transactions with neighboring users; In off-grid operation mode, the power supply priority of critical loads is dynamically adjusted based on blockchain records.

9. The energy management method based on distributed photovoltaic power according to claim 1, characterized in that, Also includes: Regularly calculate key performance indicators of system operation, including energy self-sufficiency rate, electricity cost savings rate, and grid power fluctuation rate; Based on historical operating data and performance evaluation results, the parameters of the prediction model and the optimization model are automatically tuned. The weight allocation and power command adjustment strategy for predictions at each time scale are optimized using reinforcement learning algorithms.

10. An energy management system based on distributed photovoltaic power, characterized in that, include: The data acquisition module is used to collect and standardize real-time operating data from distributed photovoltaic systems, electrochemical energy storage systems, power grids, and loads to obtain standardized real-time data. The prediction module is used to make rolling predictions of photovoltaic power output and load demand at different time scales in the future, based on the historical sequence and real-time meteorological information in the standardized real-time data and using a prediction model. The generation module is used to construct a multi-objective optimization function with the optimization objectives of system operation economy, energy self-sufficiency rate and grid impact minimization, and to generate the charging and discharging power command of the energy storage system, the power control command of the photovoltaic system and the grid connection point power plan within the current scheduling cycle based on the rolling forecast results and the standardized real-time data. The execution module is used to send the charging and discharging power command, power control command, and grid connection point power plan to the energy storage converter, photovoltaic inverter, and grid connection switchgear for execution, and monitor the actual operating status of each device during the execution process; if the deviation between the actual operating status and the corresponding predicted value exceeds a set threshold, the charging and discharging power command of the energy storage system, the power control command of the photovoltaic system, and the grid connection point power plan for the current scheduling cycle are regenerated. The monitoring module is used to perform islanding detection, harmonic analysis, and equipment status diagnosis in parallel throughout the data acquisition, instruction generation, and execution process. When the islanding detection identifies a grid fault, it controls the grid-connected switching equipment to disconnect the grid connection according to the power control instruction and the grid connection point power plan, and controls the energy storage converter to switch to off-grid operation mode according to the charging and discharging power instruction to supply power to critical loads.

Images