V2G bidirectional energy management system and method for distributed wind-solar-storage-charging pile clusters

Abstract

This invention relates to the fields of new energy power generation and electric vehicle charging technology, and particularly to a V2G bidirectional energy management system and method for distributed wind, solar, and energy storage charging pile clusters, comprising: a central dispatch server and multiple edge control nodes. A data acquisition and prediction module generates predicted data; a global optimization module generates and issues day-ahead global power commands based on this data; a local real-time control module generates and executes real-time power adjustment commands based on model predictive control, targeting the tracking commands; a communication monitoring and switching module switches to a network outage mode during communication interruptions, independently controls and caches data based on a local rule base; a network outage resumption and synchronization module uploads cached data and synchronizes key states through a consensus algorithm after communication is restored; a data fusion and recovery module performs data fusion and verification based on a global timestamp, updates the global database, and issues confirmation commands. This invention achieves autonomous operation during communication interruptions and data consistency after recovery.

Description

Technical Field

[0001] This invention relates to the field of new energy power generation and electric vehicle charging technology, and in particular to a V2G bidirectional energy management system and method for distributed wind, solar and energy storage charging pile groups. Background Technology

[0002] With the rapid development of distributed photovoltaic (PV) power generation and the continuous growth of electric vehicle ownership, distributed wind-solar-storage-charging network (FSN) systems, integrating PV power generation, energy storage devices, and V2G bidirectional charging piles, are gradually becoming an important component of urban smart energy infrastructure. These systems effectively increase the proportion of clean energy consumption and reduce the impact of electric vehicle charging on the power grid by organically combining renewable energy power generation with the charging needs of electric vehicles. In actual operation, these systems typically employ a central controller to uniformly schedule each charging pile and energy storage unit to achieve global energy optimization.

[0003] However, the existing energy management solutions for distributed wind, solar, energy storage and charging pile clusters rely excessively on stable communication between the central controller and each edge node. When the communication network fails or is interrupted, the edge nodes cannot receive central dispatch instructions and often have to switch to a simple local protection mode or shut down directly. This results in the photovoltaic power generation not being effectively absorbed during the grid outage, the energy storage unit not being able to participate in regulation, and the V2G bidirectional function failing, which seriously affects the reliability of system operation and the utilization rate of renewable energy. Summary of the Invention

[0004] To overcome the above shortcomings, this invention provides a V2G bidirectional energy management system and method for distributed wind, solar, energy storage and charging pile groups, aiming to improve the problems of distributed wind, solar, energy storage and charging pile groups being unable to operate effectively when communication is interrupted, data loss during network outages, and inconsistency in the status of multiple nodes after recovery.

[0005] In a first aspect, the present invention provides the following technical solution: a V2G bidirectional energy management system for a distributed wind-solar-storage-charging pile cluster. The system includes a central dispatch server and multiple edge control nodes. Each edge control node is connected to a wind-solar-storage input unit, an energy storage unit, and a charging pile cluster within its jurisdiction. The central dispatch server is communicatively connected to the multiple edge control nodes. The system includes the following modules:

[0006] The data acquisition and forecasting module is used to acquire weather forecast data and historical load data, and generate photovoltaic output forecast data and charging load forecast data.

[0007] The global optimization module is used to generate day-ahead global power commands based on predicted data, including energy storage charging and discharging plans and V2G scheduling plans, and to issue them to each edge control node.

[0008] The local real-time control module is used to collect the status of the equipment under its jurisdiction in real time, with the goal of tracking the global power command before the day, and generates and executes real-time power adjustment commands for energy storage and charging piles based on model predictive control.

[0009] The communication monitoring and switching module is used to monitor the communication status with the central dispatch server. When communication is interrupted, it switches to the network disconnection mode, independently executes charging and discharging control based on the local rule base, and stores the data with timestamps in the local cache.

[0010] The network outage resume and synchronization module is used to upload locally cached data after communication is restored, and to synchronize key status data with neighboring nodes through a distributed consensus algorithm;

[0011] The data fusion and recovery module is used to perform data fusion and status verification based on uploaded data and global time stamps, update the global status database, and send status confirmation instructions to the corresponding edge control nodes.

[0012] Preferably, in the data acquisition and prediction module, the steps of generating photovoltaic power output prediction data and charging load prediction data specifically include:

[0013] Obtain meteorological forecast data, which includes at least predicted irradiance and predicted ambient temperature.

[0014] Obtain the historical operation database and extract historical photovoltaic power output data and corresponding historical meteorological data, historical charging load data and corresponding historical time characteristic data;

[0015] The predicted irradiance and ambient temperature are input into the photovoltaic output prediction model based on a long short-term memory neural network to generate a photovoltaic output prediction curve.

[0016] The historical charging load data and the time characteristic data of the period to be predicted are input into the charging load prediction model based on time series analysis to generate a charging load prediction curve.

[0017] The photovoltaic power output prediction curve and the charging load prediction curve are aligned along the time axis to form a comprehensive prediction dataset.

[0018] Preferably, in the global optimization module, the step of generating a day-ahead global power command that includes an energy storage charging and discharging plan and a V2G scheduling plan specifically includes:

[0019] Obtain a comprehensive forecast dataset, which includes time-axis aligned photovoltaic power output forecast curves and charging load forecast curves;

[0020] Obtain system configuration parameters, which include at least the rated capacity of the distribution transformer, the SOC of the energy storage, and the power.

[0021] Obtain V2G vehicle dispatchability assessment data reported by each edge control node;

[0022] Based on the photovoltaic output forecast curve and the charging load forecast curve, determine the net load forecast values ​​for each time period of the next day;

[0023] With the goal of minimizing operating costs, and constrained by the capacity of the distribution transformer, the SOC and power of energy storage, and the dispatchable power of V2G, a mixed integer linear programming model is constructed and solved in conjunction with the net load forecast value to obtain the planned values ​​of energy storage charging and discharging power and V2G dispatching plan values ​​for each time period of the next day.

[0024] The energy storage charging and discharging power plan value and the V2G scheduling plan value are aligned according to the time axis to generate a day-ahead global power command and send it to each edge control node.

[0025] Preferably, in the local real-time control module, the step of generating and executing real-time power adjustment commands for energy storage and charging piles based on model predictive control specifically includes:

[0026] Obtain the day-ahead global power command, which includes the energy storage charging and discharging power plan value and the V2G scheduling plan value for the current time period;

[0027] Real-time data collection of the current power generation of the wind, solar and energy storage input units under its jurisdiction, the current SOC and current chargeable / discharge limit of the energy storage units, and the current charging / discharging power and connection status of each charging pile;

[0028] The current time period energy storage charging and discharging power plan value and V2G scheduling plan value are used as tracking targets, the current power generation power and current charging and discharging power are used as disturbance inputs, the current SOC and current charging and discharging power limit value and the connection status of each charging pile are used as constraint boundaries, and a model prediction control model is constructed with a preset time as the control period.

[0029] The predictive control model is solved by rolling to obtain the power adjustment command sequence of the energy storage unit and each charging pile in the future control time domain;

[0030] The first instruction value in the power adjustment instruction sequence is extracted as the real-time power adjustment instruction for the current period and sent to the energy storage units and charging piles under its jurisdiction for execution.

[0031] Preferably, in the communication monitoring and switching module, the step of switching to the network outage mode when communication is interrupted specifically includes:

[0032] The system periodically sends heartbeat signals to the central scheduling server and monitors whether a heartbeat response is received within a preset timeout period.

[0033] When the number of consecutive times no heartbeat response is received reaches a preset threshold, communication is determined to be interrupted, and the system automatically switches to the offline independent control mode.

[0034] In the offline independent control mode, the execution of the daytime global power command is stopped, and the charging and discharging control strategy based on the local rule base is switched.

[0035] The system collects real-time operating status data of the equipment under its jurisdiction, generates local control commands for energy storage units and charging piles based on the local rule base, and executes them.

[0036] The running status data and local control commands, with timestamps, are written into the local cache database of the edge control node.

[0037] Preferably, the step of generating and executing local control commands for the energy storage unit and charging pile based on the local rule base specifically includes:

[0038] Real-time data collection includes the current load rate of the distribution transformers under its jurisdiction, the current power generation of the wind, solar and energy storage input units, the current SOC of the energy storage units, and the current power and connection status of each charging pile.

[0039] The current load rate of the distribution transformer is compared with a preset overload threshold. When the current load rate exceeds the first overload threshold, a power limiting command is generated to reduce the charging power of the charging piles under its jurisdiction by a preset ratio.

[0040] When the current load rate exceeds the second overload threshold, an energy storage discharge command is generated to control the energy storage unit to discharge at the maximum allowable power to support the distribution transformer.

[0041] The current power generation of the wind-solar-storage input unit is compared with the current total charging power of the charging piles under its jurisdiction. When the current power generation is greater than the current total charging power and the difference exceeds a preset threshold, an energy storage charging command is generated to control the energy storage unit to absorb excess photovoltaic power.

[0042] The current SOC of the energy storage unit is compared with a preset lower limit threshold for SOC. When the current SOC is lower than the lower limit threshold, a charging priority command is generated to restrict the energy storage unit from discharging and prioritize its charging to a safe range.

[0043] The generated power limiting command, energy storage discharge command, energy storage charging command, or charging priority command will be sent to the energy storage units and charging piles under its jurisdiction for execution.

[0044] Preferably, in the network outage resumption and synchronization module, the steps of uploading locally cached data after communication is restored and synchronizing key state data with neighboring nodes through a distributed consensus algorithm specifically include:

[0045] The communication status with the central dispatch server is continuously monitored. When a heartbeat response is received from the central dispatch server or data is successfully sent to the central dispatch server, it is determined that communication has been restored.

[0046] After communication is restored, read the running data stored in the local cache database during the network outage;

[0047] The read local cache data is packaged in chronological order and uploaded to the central scheduling server through a breakpoint resume mechanism;

[0048] Establish temporary communication connections with adjacent edge control nodes, and elect a temporary leader node through a distributed consensus algorithm;

[0049] Through the coordination of the temporary leader node, each adjacent edge control node exchanges critical status data during the network outage;

[0050] Based on the exchanged data, each edge control node runs a preset consistency synchronization algorithm to ensure that each node reaches a consistent value for the key state data.

[0051] After data synchronization is completed, each edge control node reports the synchronization completion status to the central scheduling server.

[0052] Preferably, in the data fusion and recovery module, the steps of performing data fusion and status verification based on uploaded data and global time stamps, updating the global status database, and issuing status confirmation commands to the corresponding edge control nodes specifically include:

[0053] Receive locally cached data uploaded by each edge control node after communication is restored;

[0054] Extract the global time stamp from the data uploaded by each edge control node, and sort the data from different edge control nodes according to the time sequence.

[0055] Data from different edge control nodes at the same time are correlated and matched to identify and mark conflicting data records.

[0056] For data records marked as conflicting, valid data is selected based on the global timestamp and the latest timestamp principle is used to discard expired or duplicate data.

[0057] The merged data is written into the system's global status database to update the system's operating status records that were missing during the network outage.

[0058] Send status confirmation commands to each edge control node to notify them that the uploaded data has been successfully received and the fusion process has been completed;

[0059] The edge control node that receives the status confirmation instruction clears the confirmed historical data in its local cache database to free up storage space.

[0060] Secondly, the present invention provides the following technical solution: a V2G bidirectional energy management method for distributed wind-solar-storage-charging pile clusters, the method comprising:

[0061] Acquire weather forecast data and historical load data to generate photovoltaic power output forecast data and charging load forecast data;

[0062] Based on the predicted data, a day-ahead global power command containing energy storage charging and discharging plans and V2G scheduling plans is generated and sent to each edge control node;

[0063] The system collects the status of the equipment under its jurisdiction in real time, with the goal of tracking the global power command in advance. Based on model predictive control, it generates and executes real-time power adjustment commands for energy storage and charging piles.

[0064] Monitor the communication status with the central dispatch server, switch to offline mode when communication is interrupted, independently execute charging and discharging control based on the local rule base, and store the data with timestamps in the local cache;

[0065] After communication is restored, locally cached data is uploaded, and key status data is synchronized with neighboring nodes through a distributed consensus algorithm.

[0066] Data fusion and status verification are performed based on uploaded data and global time stamps, the global status database is updated, and status confirmation instructions are sent to the corresponding edge control nodes.

[0067] The present invention has the following beneficial effects:

[0068] 1. In this invention, by setting up communication status monitoring and switching modules at each edge control node, when a communication interruption with the central dispatch server is detected, the system automatically switches to the independent control mode when the network is disconnected. The charging and discharging control is executed independently based on the local rule base, and the running data is stored in the local cache with timestamps. This avoids system downtime or loss of control due to communication failures and ensures the continuous consumption of wind and solar resources and the normal operation of charging services during the network outage.

[0069] 2. In this invention, the local cached data is uploaded after communication is restored through the network interruption resume and synchronization module. The key status data during the network interruption is synchronized with the adjacent edge control nodes through a distributed consensus algorithm. Combined with the data fusion and recovery module of the central scheduling server, data verification and fusion are performed based on the global time stamp. This effectively solves the problem of multi-node data conflict and ensures the accuracy and continuity of the system's global status.

[0070] 3. In this invention, the global optimization module generates a day-ahead global power command that includes an energy storage charging and discharging plan and a V2G scheduling plan based on predicted data. Combined with the local real-time control module, a real-time power adjustment command is generated and executed based on model predictive control. This achieves efficient bidirectional allocation of wind, solar and energy storage resources within the charging pile group, maximizes the absorption of photovoltaic power generation, reduces the cost of purchasing electricity from the grid, and improves the economic efficiency of system operation. Attached Figure Description

[0071] Figure 1 This is a schematic diagram of the architecture of the V2G bidirectional energy management system for the distributed wind-solar-storage-charging pile cluster proposed in this invention.

[0072] Figure 2 This is a flowchart illustrating the V2G bidirectional energy management method for distributed wind-solar-storage-charging pile clusters proposed in this invention. Detailed Implementation

[0073] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0074] Example 1:

[0075] In the first embodiment of the present invention, the present invention provides a V2G bidirectional energy management system for a distributed wind-solar-storage-charging pile cluster, such as... Figure 1 As shown, the system includes a central dispatch server and multiple edge control nodes. The edge control nodes are connected to the wind-solar-storage input units, energy storage units, and charging pile groups within their respective jurisdictions. The central dispatch server communicates with the multiple edge control nodes and includes the following modules:

[0076] The data acquisition and forecasting module is used to acquire weather forecast data and historical load data, and generate photovoltaic power output forecast data and charging load forecast data.

[0077] Specifically, firstly, the data acquisition and forecasting module obtains weather forecast data for the target area through a meteorological interface service. The weather forecast data includes at least the predicted irradiance, ambient temperature, and wind speed every 15 minutes for the next 72 hours. Simultaneously, the module extracts historical photovoltaic power output data and corresponding historical meteorological data for the past 90 days, as well as historical charging load data and corresponding historical time characteristic data, from the local historical operating database. The historical time characteristic data includes information such as time of day, week type, whether it is a holiday, and season.

[0078] The construction and prediction process of the photovoltaic (PV) power output prediction model is as follows: A long short-term memory (LSTM) neural network is used to construct the PV power output prediction model. This neural network model contains one input layer, two LSTM hidden layers, and one fully connected output layer. Each LSTM hidden layer contains 128 memory units and uses tanh as the activation function. The feature vector received by the input layer includes predicted irradiance, predicted ambient temperature, predicted wind speed, and historical PV power output values. During model training, historical meteorological data is used as input, and the corresponding historical PV power output data is used as output. The Adam optimizer is used for training, and the loss function is mean squared error. After training, the real-time predicted irradiance, predicted ambient temperature, and predicted wind speed values ​​are input into the trained PV power output prediction model. The model outputs PV power output prediction values ​​every 15 minutes for the next 72 hours, forming a PV power output prediction curve. .

[0079] The construction and prediction process of the charging load forecasting model is as follows: A time series analysis algorithm based on a differential integrated moving average autoregressive model is used to construct the charging load forecasting model. First, the stationarity and seasonality decomposition of historical charging load data are performed to determine the model parameters p, d, and q. The model expression is:

[0080] ;

[0081] in, Let L be the charging load value at time t, and L be the hysteresis operator. These are the autoregressive coefficients. Here, d is the moving average coefficient, s is the difference order, and D is the seasonal difference order. The data is a white noise sequence. During model training, historical charging load data and corresponding historical time feature data are used as input, and maximum likelihood estimation is used to solve for the parameters. After training, the time feature data of the period to be predicted are input into the trained charging load prediction model. The model outputs the predicted charging load values ​​at 15-minute intervals for the next 72 hours, forming a charging load prediction curve. .

[0082] Finally, the data acquisition and prediction module will generate the photovoltaic output prediction curve. Align the two curves with the charging load prediction curve P_ev(t) on the time axis to ensure that they have the same time resolution and time span, thus forming a comprehensive prediction dataset. This is provided for use by the global optimization module.

[0083] The global optimization module is used to generate day-ahead global power commands based on predicted data, including energy storage charging and discharging plans and V2G scheduling plans, and then distribute them to each edge control node.

[0084] Specifically, the global optimization module is deployed in the central dispatch server. It generates day-ahead global power commands, including energy storage charging / discharging plans and V2G dispatch plans, based on forecast data and distributes them to each edge control node. The specific implementation process is as follows: The global optimization module first obtains the comprehensive forecast dataset from the data acquisition and forecast module. This comprehensive forecast dataset contains time-aligned photovoltaic power output forecast curves. and charging load forecast curve , where t represents discrete time points at 15-minute intervals, and t=1,2,…,96 correspond to different time periods in the next 24 hours.

[0085] The global optimization module obtains system configuration parameters. These parameters include at least the rated capacity of the distribution transformer. Rated power of energy storage unit and rated capacity SOC operating limit of energy storage units and lower limit of operation In addition, the global optimization module obtains V2G vehicle dispatchability assessment data reported by each edge control node. This assessment data includes the number of vehicles that can participate in discharge and the dispatchable power margin for each time period of the following day. (t).

[0086] The global optimization module is based on the photovoltaic output prediction curve. and charging load forecast curve Determine the net load forecast for each time period of the following day. The net load forecast is calculated using the following formula:

[0087] ;

[0088] in Let be the net load forecast at time t. Let t be the predicted charging load value. Let be the predicted photovoltaic power output at time t. A positive value indicates that the system needs to purchase electricity from the grid or supplement the power deficit through energy storage discharge. A negative value indicates that there is excess photovoltaic power generation that needs to be absorbed by energy storage charging or V2G vehicle charging.

[0089] The global optimization module constructs an objective function with the goal of minimizing operating costs. Operating costs include the cost of purchasing electricity from the grid, the cost of energy storage charging and discharging losses, and the cost of V2G discharge compensation. The objective function expression is as follows:

[0090] ;

[0091] Where t is the time period number, ranging from 1 to 96. The time interval for each period is 0.25 hours. Let t be the grid purchase price of electricity. Let t be the power exchanged between the system and the grid. A positive value indicates that the system purchases electricity from the grid, while a negative value indicates that the system sells electricity to the grid. The unit loss cost coefficient for energy storage charging and discharging. Let t be the energy storage charging power. Let be the energy storage discharge power at time t. Let be the unit compensation cost coefficient for V2G discharge at time t. Let t be the total discharge power of the V2G vehicle.

[0092] The global optimization module constructs a mixed-integer linear programming model using distribution transformer capacity constraints, energy storage SOC constraints and power constraints, and V2G dispatchable power constraints as boundary conditions. The constraint expressions are as follows: Power balance constraint:

[0093] ;

[0094] in, Let t be the charging power of the V2G vehicle at time t.

[0095] Distribution transformer capacity constraints:

[0096] ;

[0097] Energy storage power constraints:

[0098] ;

[0099] ;

[0100] ;

[0101] in, and Use binary variables to represent the energy storage charging and discharging status to avoid simultaneous charging and discharging.

[0102] Energy storage SOC dynamic constraints:

[0103] ;

[0104] in The energy storage charging efficiency is set at 0.95. The energy storage discharge efficiency is set to 0.95.

[0105] Energy storage SOC upper and lower limits constraints:

[0106] ;

[0107] ;

[0108] ;

[0109] in The SOC value at the initial time. This refers to the SOC target value that must be achieved at the end of the day.

[0110] V2G power constraints:

[0111] ;

[0112] ;

[0113] ;

[0114] in, and Use binary variables to represent the charging and discharging status of V2G vehicles to avoid simultaneous charging and discharging.

[0115] The global optimization module calls the mixed-integer linear programming solver to solve the above model and obtain the planned values ​​of energy storage charging and discharging power for each time period of the next day. and V2G scheduling plan value The solver uses commercial software such as Gurobi or Cplex, and the results obtained include... , , and The value.

[0116] The global optimization module aligns the planned energy storage charging and discharging power values ​​and the V2G scheduling plan values ​​along the time axis to generate the day-ahead global power command. The current global power command is represented as:

[0117] ;

[0118] The global optimization module generates a daily global power command and sends it to each edge control node through the communication network, which serves as the tracking target for each edge control node to perform local real-time control the following day.

[0119] The local real-time control module is used to collect the status of the equipment under its jurisdiction in real time. With the goal of tracking the global power command before the day, it generates and executes real-time power adjustment commands for energy storage and charging piles based on model predictive control.

[0120] Specifically, the local real-time control module is deployed in each edge control node to collect the status of the devices under its jurisdiction in real time. With the goal of tracking the day-ahead global power command, it generates and executes real-time power adjustment commands for energy storage and charging piles based on model predictive control. The specific implementation process is as follows: The local real-time control module first obtains the day-ahead global power command from the central dispatch server. The day-ahead global power command includes the planned charging and discharging power values ​​for the energy storage during the current period. and V2G scheduling plan value , where t0 represents the start time of the current time period.

[0121] The local real-time control module collects status information from the equipment under its jurisdiction in real time. The collected information includes: the current power generation of the wind, solar, and energy storage input units. Current SOC value of the energy storage unit Current maximum rechargeable power and current maximum discharge power Current charging and discharging power of each charging station and connection status , where i represents the charging pile number, and a connection status of 1 indicates that a vehicle is connected, while 0 indicates that no vehicle is connected.

[0122] The local real-time control module uses the planned energy storage charging and discharging power for the current time period. and V2G scheduling plan value To track the target, at the current power generation capacity and the current charging and discharging power of each charging pile The disturbance input is the current SOC value. Current chargeable / dischargeable power limits and Connection status of each charging station To constrain the boundary, a preset time is used. A predictive control model is built for the control cycle. Control cycle Take 1 second.

[0123] The model predictive control (MDC) model is represented in discrete-time state-space form. The system state variables include the energy storage SOC and the power of each charging station, and the state vector is represented as follows:

[0124] ;

[0125] Where n is the number of charging piles under its jurisdiction. Control variables include energy storage capacity adjustment. and power adjustment of each charging pile The control vector is represented as:

[0126] , ;

[0127] Disturbance inputs include changes in photovoltaic power generation. and uncontrollable load changes The perturbation vector is represented as:

[0128] ;

[0129] The system state equation is expressed as:

[0130] ;

[0131] Where A is the identity matrix. and This is the coefficient matrix determined based on the system parameters.

[0132] The objective function of model predictive control aims to track the day-ahead global power command and mitigate power fluctuations at grid-connected points, and its expression is as follows:

[0133] ;

[0134] in For the prediction time domain, 5 seconds are used. To control the time domain to 1 second, The output at time k+j is predicted at time k. The reference trajectory is the target value corresponding to the global power command of the day before. To output the weight matrix, To control the weight matrix, To control the incremental weight matrix, To control the increment.

[0135] The output quantity y(k) includes the grid connection point power. (k) and the energy storage SOC value, the output equation is expressed as:

[0136] ;

[0137] Among them, C, , This is the output matrix.

[0138] The constraints include: upper and lower limits of energy storage power:

[0139] ;

[0140] Energy storage SOC upper and lower limits constraints:

[0141] ;

[0142] Charging pile power regulation rate constraint:

[0143] ;

[0144] Power limit constraints for charging piles:

[0145] ;

[0146] The local real-time control module in each control cycle The model predictive control optimization problem is solved using an internal rolling algorithm. An embedded quadratic programming solver, such as OSQP or CVXGEN, is used to solve the problem, running on the embedded processor of the edge control node. The solution yields the optimal control sequence in the future control time domain. .

[0147] The local real-time control module extracts the first control command from the optimal control sequence. This is a real-time power adjustment command for the current cycle. The command includes the power adjustment amount for the energy storage unit. and the power adjustment amount of each charging station .

[0148] The local real-time control module sends the generated real-time power adjustment commands to the energy storage units and charging piles under its jurisdiction for execution. The energy storage units then adjust the commands according to the specified parameters. Adjust the current charging and discharging power; each charging station will adjust accordingly. Adjust the current charging and discharging power. After execution, the local real-time control module waits for the next control cycle and repeats the above process of data acquisition, modeling, solving, and distribution to achieve rolling optimization control.

[0149] The communication monitoring and switching module is used to monitor the communication status with the central dispatch server. When communication is interrupted, it switches to the offline mode, independently executes charging and discharging control based on the local rule base, and stores the data with timestamps in the local cache.

[0150] Specifically, the communication monitoring and switching module is deployed in each edge control node to monitor the communication status with the central dispatch server. When communication is interrupted, it switches to a network-off mode, independently executing charging and discharging control based on a local rule base and storing the timestamped data in a local cache. The specific implementation process is as follows: The communication monitoring and switching module periodically sends heartbeat signals to the central dispatch server. The heartbeat signal sending cycle... Set a timer for 5 seconds. The module simultaneously starts a timer to monitor whether a heartbeat response is received from the central scheduling server within the preset timeout period. Timeout period. Take 15 seconds.

[0151] The communication monitoring and handover module maintains a count variable for consecutive times when no heartbeat response has been received. If no heartbeat response is received within the timeout period, Add 1; when a heartbeat response is received, Reset to zero. This refers to the number of consecutive times no heartbeat response has been received. Reaching the preset threshold Communication is considered interrupted when the time is right. Preset threshold. Take 3 times.

[0152] Upon detecting a communication interruption, the communication monitoring and switching module automatically switches to the network-disconnected independent control mode. In this mode, the module ceases executing day-ahead global power commands and switches to a charging / discharging control strategy based on a local rule base. The communication monitoring and switching module collects real-time operating status data from its managed equipment. This data includes the current load rate of the distribution transformer. Current power generation of wind, solar and energy storage input units Current SOC value of the energy storage unit Current power of each charging station and connection status .

[0153] The communication monitoring and switching module generates local control commands for the energy storage unit and charging pile based on a local rule base. The local rule base includes the following rules:

[0154] Rule 1: Overload Protection Rules for Distribution Transformers. The current load rate of the distribution transformer... Compare with a preset overload threshold. First overload threshold. Set it to 0.8, the second overload threshold. Take 0.95. When Exceed At that time, a power limit command is generated, and the power limit for each charging station is calculated according to the following formula:

[0155] ;

[0156] when Exceed At that time, an energy storage discharge command is generated to control the energy storage unit to discharge at the maximum allowable power. Discharge is used to support the distribution transformer.

[0157] Rule 2: Prioritize Photovoltaic Grid Integration. This rule prioritizes the current power generation of the wind, solar, and energy storage input units. The current total charging power of the charging piles under its jurisdiction Compare them. Calculate according to the following formula:

[0158] ;

[0159] when Greater than And the difference exceeds the preset threshold At that time, an energy storage charging command is generated to control the energy storage unit to absorb excess photovoltaic power. Preset threshold. Take 5kW. The energy storage charging power is calculated according to the following formula:

[0160] ;

[0161] Rule 3: Energy Storage SOC Safety Maintenance Rules. This rule specifies the current status of the energy storage unit. Compared with the preset SOC safety lower limit threshold Comparison. SOC safety lower limit threshold Take 0.2. When Below At this time, a charging priority command is generated to limit the discharge of the energy storage unit and prioritize its charging to a safe range. The energy storage charging power is calculated according to the following formula:

[0162] ;

[0163] in, For safety, the SOC is set to 0.5. The charging time is set to 1 hour.

[0164] The communication monitoring and switching module will send the generated power limiting command, energy storage discharge command, energy storage charging command or charging priority command to the energy storage units and charging piles under its jurisdiction for execution.

[0165] The communication monitoring and switching module writes the operating status data and local control commands, with timestamps, to the local cache database of the edge control node. The written data includes the timestamp of the acquisition time. Distribution transformer load rate Photovoltaic power generation Energy storage SOC value Power of each charging station The generated local control command types and values ​​are specified. Timestamps are accurate to at least milliseconds and use the Unix timestamp format. The local cache database is implemented using the embedded SQLite database, with a data storage capacity configured to store at least 72 hours of runtime data. When the cache space reaches its limit, a round-robin overwrite strategy is used, prioritizing the overwriting of the oldest historical data.

[0166] The network outage resume and synchronization module is used to upload locally cached data after communication is restored, and to synchronize key status data with neighboring nodes through a distributed consensus algorithm.

[0167] Specifically, the network outage resumption and synchronization module is deployed in each edge control node. It is used to upload locally cached data after communication is restored and to synchronize key status data with neighboring nodes through a distributed consensus algorithm. The specific implementation process is as follows: The network outage resumption and synchronization module continuously monitors the communication status with the central scheduling server. Monitoring methods include two aspects: first, monitoring whether a heartbeat response is received from the central scheduling server; and second, monitoring whether data is successfully sent to the central scheduling server. When a heartbeat response is received or data is successfully sent, communication is considered restored.

[0168] After communication is restored, the network interruption resumption and synchronization module reads the running data stored in the local cache database during the network interruption. The data range read starts from the time of communication interruption. Until communication is restored It contains all timestamped runtime status data and local control command execution records within the specified time period. Data is read in ascending order of timestamps to ensure the data order matches the time of occurrence.

[0169] The network interruption resume and synchronization module packages the locally cached data in chronological order and uploads it to the central dispatch server via a breakpoint resume mechanism. The breakpoint resume mechanism works as follows: the module divides the data into multiple data blocks, each 1MB in size. During upload, it records the position of the last successfully uploaded data block. If the upload is interrupted, it resumes from the point of interruption without retransmitting completed data blocks. Each data block includes a start timestamp, an end timestamp, and the data content, facilitating data integrity verification by the central dispatch server. The network interruption resume and synchronization module establishes temporary communication connections with adjacent edge control nodes. Adjacent nodes refer to other edge control nodes that are physically close or share the same power distribution transformer in terms of electrical connection. The module discovers adjacent nodes via LAN broadcast or multicast and establishes TCP communication connections.

[0170] The network outage resumption and synchronization module elects a temporary leader node through a distributed consensus algorithm. The Raft consensus algorithm is used for leader election, and the process is as follows: Each edge control node is in a follower state, with a random election timeout set between 150 and 300 milliseconds. When a follower does not receive a heartbeat message from the leader within the timeout period, it transitions to a candidate state, increments its current term number (term), and sends a vote request RPC message to all neighboring nodes. Upon receiving the vote request message, neighboring nodes vote for the candidate if its log is at least as new as their own and it has not yet voted during its current term. When a candidate receives more than half of the votes from neighboring nodes, it becomes the temporary leader node. Through the coordination of the temporary leader node, neighboring edge control nodes exchange key state data from the network outage period. This key state data includes at least the cumulative load of the distribution transformer. and regional total adjustable power margin Data exchange is implemented using a log replication mechanism: the temporary leader node encapsulates its local critical state data into log entries and sends append log RPC messages to all follower nodes. Upon receiving the message, the follower nodes write the log entries locally and return a success response.

[0171] Based on the exchanged data, each edge control node runs a preset consistency synchronization algorithm to ensure that all nodes reach a consistent value for key state data. The consistency synchronization algorithm aggregates the cumulative load of the distribution transformer according to the following formula:

[0172] ;

[0173] Where m is the number of edge control nodes participating in synchronization. This represents the cumulative load of the distribution transformer recorded during the network outage at node j. This represents the cumulative load of the distribution transformer after synchronization. For the total adjustable power margin of the area, the minimum value is taken according to the following formula to ensure safety:

[0174] ;

[0175] Each edge control node updates its local critical status data to the synchronized, consistent values. and And write it to the local state database.

[0176] After data synchronization is complete, each edge control node reports its synchronization completion status to the central dispatch server. The reported message includes the node identifier, synchronization completion timestamp, and key status data values ​​after synchronization. Once the central dispatch server receives the synchronization completion status from all relevant nodes, it confirms that the state synchronization process for that area is complete.

[0177] The data fusion and recovery module is used to perform data fusion and status verification based on uploaded data and global time stamps, update the global status database, and send status confirmation instructions to the corresponding edge control nodes.

[0178] Specifically, the data fusion and recovery module is deployed in the central scheduling server. It performs data fusion and status verification based on uploaded data and global time stamps, updates the global status database, and sends status confirmation commands to the corresponding edge control nodes. The specific implementation process is as follows: The data fusion and recovery module receives locally cached data uploaded by each edge control node after communication is restored. The received data includes all operational status data and local control command execution records recorded by each edge control node during the network outage. Each data record contains a global time stamp generated by the edge control node, accurate to the millisecond level. The data fusion and recovery module extracts the global time stamp from the data uploaded by each edge control node and sorts the data from different edge control nodes according to chronological order. The sorting is based on the global time stamp as the primary key, using ascending order to form a queue of data to be processed arranged along a timeline.

[0179] The data fusion and recovery module correlates and matches data from different edge control nodes at the same time. The correlation and matching are based on global timestamps and device identifiers, grouping data records with the same timestamp and describing the same device status together. For each group, the module identifies and marks conflicting data records. Conflict types include data conflicts with inconsistent values, redundant data due to communication delays, and timing discrepancies due to node clock drift. For data records marked as conflicting, the data fusion and recovery module selects valid data based on the latest timestamp, using the global timestamp as the benchmark. The data selection rule is as follows:

[0180] ;

[0181] Where t is time, and S(t) is the set of data records from different edge control nodes at time t. For the data record of the j-th node at time t, This is the global time stamp of the data record. Select valid data records. The module discards expired and duplicate data, retaining only the valid data records with the latest timestamp.

[0182] To address timing discrepancies caused by node clock drift, the data fusion and recovery module employs a clock correction algorithm based on NTP time synchronization. The module compares the timestamps of data uploaded by each node with the reference time of the central scheduling server to calculate the time deviation for each node. The time deviation is calculated using the following formula:

[0183] ;

[0184] Where N is the number of data points used for calibration. Let k be the timestamp of the k-th data record in the j-th node. This is the corresponding reception time recorded by the central scheduling server. The module corrects the timestamps of subsequent data uploaded by each node based on the calculated time deviation. The corrected timestamps are... Calculate according to the following formula:

[0185] ;

[0186] The data fusion and recovery module writes the fused data into the system's global status database. The written data includes timestamps, device identifiers, device types, operating status parameter values, control command types, and command values. Through fusion processing, the module completely supplements the system operating status records missing during the network outage, restoring the global status database to a state consistent with the actual situation.

[0187] The data fusion and recovery module sends status confirmation commands to each edge control node. The status confirmation command includes the start and end time range of the data uploaded by the edge control node, in the following format: This indicates that the data uploaded within this time period has been successfully received and processed.

[0188] The edge control node that receives the status confirmation command, according to the time range in the command... Clear the historical data for the corresponding time period from the local cache database. After clearing, the edge control node releases the corresponding storage space, preparing it for caching data in case of subsequent network outages. If the edge control node receives multiple status confirmation commands, it only clears the data for the corresponding time period each time, retaining the unconfirmed historical data to be re-uploaded after the next communication is restored.

[0189] Example 2:

[0190] Existing energy management solutions for distributed wind-solar-storage-charging pile clusters rely excessively on stable communication between the central controller and each edge node. When the communication network fails or is interrupted, the edge nodes cannot receive central dispatch commands and often have to switch to a simple local protection mode or shut down directly. This results in ineffective consumption of photovoltaic power generation, inability of energy storage units to participate in regulation, and failure of V2G bidirectional functionality during the outage. Furthermore, after communication is restored, the operational data of each node during the outage cannot be effectively synchronized and integrated, causing inconsistencies in the overall system state and severely impacting the reliability of system operation, renewable energy utilization, and data integrity. To address these issues, this invention provides a V2G bidirectional energy management method for distributed wind-solar-storage-charging pile clusters, the structure of which is as follows: Figure 2 As shown. The specific implementation process of this method is as follows:

[0191] Specifically, the process begins by acquiring weather forecast data and historical load data to generate photovoltaic output forecast data and charging load forecast data. Then, based on the forecast data, a day-ahead global power command, including energy storage charging and discharging plans and V2G scheduling plans, is generated and sent to each edge control node. Each edge control node collects the status of its managed equipment in real time and, with the day-ahead global power command as the target, generates and executes real-time power adjustment commands for energy storage and charging piles based on model predictive control. Simultaneously, it continuously monitors the communication status with the central dispatch server and automatically switches to a network outage mode when communication is interrupted. It independently executes charging and discharging control based on a local rule base and stores the running data with timestamps in a local cache. After communication is restored, the local cache data is uploaded, and key status data is synchronized with neighboring nodes through a distributed consensus algorithm. Finally, the central dispatch server performs data fusion and status verification based on the uploaded data and global timestamps, updates the global status database, and sends status confirmation commands to the corresponding edge control nodes. This achieves a complete energy management closed loop from prediction optimization to real-time control to network outage autonomy and recovery.

[0192] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A V2G bidirectional energy management system for a distributed wind-solar-storage-charging pile cluster, the system comprising a central dispatch server and multiple edge control nodes, wherein each edge control node is connected to a wind-solar-storage input unit, an energy storage unit, and a charging pile cluster within its jurisdiction, and the central dispatch server is communicatively connected to the multiple edge control nodes, characterized in that, Includes the following modules: The data acquisition and forecasting module is used to acquire weather forecast data and historical load data, and generate photovoltaic output forecast data and charging load forecast data. The global optimization module is used to generate day-ahead global power commands based on predicted data, including energy storage charging and discharging plans and V2G scheduling plans, and to issue them to each edge control node. The local real-time control module is used to collect the status of the equipment under its jurisdiction in real time. With the goal of tracking the global power command before the day, it generates and executes real-time power adjustment commands for energy storage and charging piles based on model predictive control. The communication monitoring and switching module is used to monitor the communication status with the central dispatch server. When communication is interrupted, it switches to the network disconnection mode, independently executes charging and discharging control based on the local rule base, and stores the data with timestamps in the local cache. The network outage resume and synchronization module is used to upload locally cached data after communication is restored, and to synchronize key status data with neighboring nodes through a distributed consensus algorithm; The data fusion and recovery module is used to perform data fusion and status verification based on uploaded data and global time stamps, update the global status database, and send status confirmation instructions to the corresponding edge control nodes.

2. The V2G bidirectional energy management system for distributed wind-solar-storage charging pile clusters according to claim 1, characterized in that, In the data acquisition and prediction module, the steps for generating photovoltaic power output prediction data and charging load prediction data specifically include: Obtain meteorological forecast data, which includes at least predicted irradiance and predicted ambient temperature. Obtain the historical operation database and extract historical photovoltaic power output data and corresponding historical meteorological data, historical charging load data and corresponding historical time characteristic data; The predicted irradiance and ambient temperature are input into the photovoltaic output prediction model based on a long short-term memory neural network to generate a photovoltaic output prediction curve. The historical charging load data and the time characteristic data of the period to be predicted are input into the charging load prediction model based on time series analysis to generate a charging load prediction curve. The photovoltaic power output prediction curve and the charging load prediction curve are aligned along the time axis to form a comprehensive prediction dataset.

3. The V2G bidirectional energy management system for distributed wind-solar-storage charging pile clusters according to claim 2, characterized in that, In the global optimization module, the step of generating a day-ahead global power command that includes an energy storage charging and discharging plan and a V2G scheduling plan specifically includes: Obtain a comprehensive forecast dataset, which includes time-axis aligned photovoltaic power output forecast curves and charging load forecast curves; Obtain system configuration parameters, which include at least the rated capacity of the distribution transformer, the SOC of the energy storage, and the power. Obtain V2G vehicle dispatchability assessment data reported by each edge control node; Based on the photovoltaic output forecast curve and the charging load forecast curve, determine the net load forecast values ​​for each time period of the next day; With the goal of minimizing operating costs, and constrained by the capacity of the distribution transformer, the SOC and power of energy storage, and the dispatchable power of V2G, a mixed integer linear programming model is constructed and solved in conjunction with the net load forecast value to obtain the planned values ​​of energy storage charging and discharging power and V2G dispatching plan values ​​for each time period of the next day. The planned energy storage charging and discharging power values ​​and the V2G scheduling plan values ​​are aligned along the time axis to generate a day-ahead global power command and send it to each edge control node.

4. The V2G bidirectional energy management system for distributed wind-solar-storage charging pile clusters according to claim 1, characterized in that, In the local real-time control module, the steps of generating and executing real-time power adjustment commands for energy storage and charging piles based on model predictive control specifically include: Obtain the day-ahead global power command, which includes the energy storage charging and discharging power plan value and the V2G scheduling plan value for the current time period; Real-time data collection of the current power generation of the wind, solar and energy storage input units under its jurisdiction, the current SOC and current chargeable / discharge limit of the energy storage units, and the current charging / discharging power and connection status of each charging pile; The current time period energy storage charging and discharging power plan value and V2G scheduling plan value are used as tracking targets, the current power generation power and current charging and discharging power are used as disturbance inputs, the current SOC and current charging and discharging power limit value and the connection status of each charging pile are used as constraint boundaries, and a model prediction control model is constructed with a preset time as the control period. The predictive control model is solved by rolling to obtain the power adjustment command sequence of the energy storage unit and each charging pile in the future control time domain; The first instruction value in the power adjustment instruction sequence is extracted as the real-time power adjustment instruction for the current period and sent to the energy storage units and charging piles under its jurisdiction for execution.

5. The V2G bidirectional energy management system for distributed wind-solar-storage charging pile clusters according to claim 1, characterized in that, In the communication monitoring and switching module, the steps for switching to the network outage mode when communication is interrupted specifically include: The system periodically sends heartbeat signals to the central scheduling server and monitors whether a heartbeat response is received within a preset timeout period. When the number of consecutive times no heartbeat response is received reaches a preset threshold, communication is determined to be interrupted, and the system automatically switches to the offline independent control mode. In the offline independent control mode, the execution of the daytime global power command is stopped, and the charging and discharging control strategy based on the local rule base is switched. The system collects real-time operating status data of the equipment under its jurisdiction, generates local control commands for energy storage units and charging piles based on the local rule base, and executes them. The running status data and local control commands, with timestamps, are written into the local cache database of the edge control node.

6. The V2G bidirectional energy management system for distributed wind-solar-storage charging pile clusters according to claim 5, characterized in that, The step of generating and executing local control commands for energy storage units and charging piles based on the local rule base specifically includes: Real-time data collection includes the current load rate of the distribution transformers under its jurisdiction, the current power generation of the wind, solar and energy storage input units, the current SOC of the energy storage units, and the current power and connection status of each charging pile. The current load rate of the distribution transformer is compared with a preset overload threshold. When the current load rate exceeds the first overload threshold, a power limiting command is generated to reduce the charging power of the charging piles under its jurisdiction by a preset ratio. When the current load rate exceeds the second overload threshold, an energy storage discharge command is generated to control the energy storage unit to discharge at the maximum allowable power to support the distribution transformer. The current power generation of the wind-solar-storage input unit is compared with the current total charging power of the charging piles under its jurisdiction. When the current power generation is greater than the current total charging power and the difference exceeds a preset threshold, an energy storage charging command is generated to control the energy storage unit to absorb excess photovoltaic power. The current SOC of the energy storage unit is compared with a preset lower limit threshold for SOC. When the current SOC is lower than the lower limit threshold, a charging priority command is generated to restrict the energy storage unit from discharging and prioritize its charging to a safe range. The generated power limiting command, energy storage discharge command, energy storage charging command, or charging priority command will be sent to the energy storage units and charging piles under its jurisdiction for execution.

7. The V2G bidirectional energy management system for distributed wind-solar-storage charging pile clusters according to claim 1, characterized in that, In the network outage resumption and synchronization module, the steps of uploading locally cached data after communication is restored and synchronizing key state data with neighboring nodes through a distributed consensus algorithm specifically include: The communication status with the central dispatch server is continuously monitored. When a heartbeat response is received from the central dispatch server or data is successfully sent to the central dispatch server, it is determined that communication has been restored. After communication is restored, read the running data stored in the local cache database during the network outage; The read local cache data is packaged in chronological order and uploaded to the central scheduling server through a breakpoint resume mechanism; Establish temporary communication connections with adjacent edge control nodes, and elect a temporary leader node through a distributed consensus algorithm; Through the coordination of the temporary leader node, each adjacent edge control node exchanges critical status data during the network outage; Based on the exchanged data, each edge control node runs a preset consistency synchronization algorithm to ensure that each node reaches a consistent value for the key state data. After data synchronization is completed, each edge control node reports the synchronization completion status to the central scheduling server.

8. The V2G bidirectional energy management system for distributed wind-solar-storage charging pile clusters according to claim 1, characterized in that, In the data fusion and recovery module, the steps of performing data fusion and status verification based on uploaded data and global time stamps, updating the global status database, and issuing status confirmation commands to the corresponding edge control nodes specifically include: Receive locally cached data uploaded by each edge control node after communication is restored; Extract the global time stamp from the data uploaded by each edge control node, and sort the data from different edge control nodes according to the time sequence. Data from different edge control nodes at the same time are correlated and matched to identify and mark conflicting data records. For data records marked as conflicting, valid data is selected based on the global timestamp and the latest timestamp principle is used to discard expired or duplicate data. The merged data is written into the system's global status database to update the system's operating status records that were missing during the network outage. Send status confirmation commands to each edge control node to notify them that the uploaded data has been successfully received and the fusion process has been completed; The edge control node that receives the status confirmation instruction clears the confirmed historical data in its local cache database to free up storage space.

9. A V2G bidirectional energy management method for distributed wind-solar-storage charging pile clusters, characterized in that, The method applied to the V2G bidirectional energy management system of the distributed wind-solar-storage-charging pile group according to any one of claims 1-8 includes: Acquire weather forecast data and historical load data to generate photovoltaic power output forecast data and charging load forecast data; Based on the predicted data, a day-ahead global power command containing energy storage charging and discharging plans and V2G scheduling plans is generated and sent to each edge control node; The system collects the status of the equipment under its jurisdiction in real time, with the goal of tracking the global power command in advance. Based on model predictive control, it generates and executes real-time power adjustment commands for energy storage and charging piles. Monitor the communication status with the central dispatch server, switch to offline mode when communication is interrupted, independently execute charging and discharging control based on the local rule base, and store the data with timestamps in the local cache; After communication is restored, locally cached data is uploaded, and key status data is synchronized with neighboring nodes through a distributed consensus algorithm. Data fusion and status verification are performed based on uploaded data and global time stamps, the global status database is updated, and status confirmation instructions are sent to the corresponding edge control nodes.

Images