Electric power regulation and control system and device based on photovoltaic intelligent micro breaker

Abstract

The invention relates to the technical field of power regulation and control systems, and discloses a power regulation and control system and device based on a photovoltaic intelligent micro circuit breaker, and the device comprises a data collection unit which is used for obtaining the operation parameters of a power grid, and collecting the energy storage equipment data, power generation data and load data of distributed photovoltaic users based on the photovoltaic intelligent micro circuit breaker; a strategy unit configured with a deep reinforcement learning model and used for generating a power generation strategy and an energy storage strategy of each distributed photovoltaic user; the virtual power storage unit is used for recording and updating the energy storage equipment data, the power generation data, the load data and the power grid operation parameters; and the control instruction unit is used for generating a control instruction of each photovoltaic user based on the power generation strategy and the energy storage strategy and sending the control instruction to the corresponding photovoltaic intelligent micro circuit breaker. According to the invention, the distributed energy storage equipment can be used for coordinating photovoltaic power generation and power utilization requirements of users, and the energy utilization efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power control systems, and discloses a power control system and device based on a photovoltaic intelligent micro-breaker. Background Art

[0002] Currently, the construction of smart grids enables real-time monitoring and control of distributed photovoltaic systems. However, distributed photovoltaic systems still face numerous challenges in power regulation. Due to the random and intermittent nature of photovoltaic power generation, its output power can fluctuate significantly due to factors such as sunlight intensity and weather changes. If the total power uploaded to the grid by distributed photovoltaic users is too high, the voltage connected to the grid may be excessive, even exceeding the rated range of the grid transformer, causing damage to the transformer and compromising the safe and stable operation of the grid.

[0003] Currently, there are significant deficiencies in the management of energy storage devices. On the one hand, there is a lack of rational control strategies for the charging and discharging of energy storage devices, which makes overcharging and discharging prone to occur. On the other hand, insufficient consideration is given to the coordinated optimization of energy storage devices with photovoltaic power generation and user electricity consumption, which prevents them from maximizing their advantages and improving energy efficiency. For example, when the grid load is low and the grid power generation is sufficient, or when user power consumption is high and the grid power generation is insufficient, the discharge of energy storage devices cannot be effectively utilized to uniformly reduce peak loads and fill valleys, rendering distributed energy storage devices ineffective in contributing to the stable operation of the grid.

[0004] For example, the existing Chinese patent with authorization announcement number CN116780660B discloses a hierarchical collaborative control method and system for distributed photovoltaics, which includes: analyzing whether the total power generation of all distributed photovoltaics in the current control area is greater than the total load in the current control area; if not, generating a control instruction; when triggering the control program, giving priority to adjacent control areas for secondment, analyzing the difficulty of secondment of multiple adjacent control areas, selecting adjacent control areas with high response enthusiasm, fast response and easy to be called, comprehensively considering the surplus electricity of each adjacent control area, allocating the borrowed electricity to the selected adjacent control area, generating an allocation control strategy, and executing the control of electricity call according to the allocation control strategy, realizing real-time monitoring, efficient coordination and orderly management of distributed photovoltaic systems, forming a management mode of distributed photovoltaic group control and group adjustment, coordinating controllable resources for large-scale scheduling over a long period of time, and improving the stable operation of the power grid.

[0005] For example, a Chinese patent application with publication number CN118487327A discloses a method and system for regional distributed photovoltaic aggregation control in a power system. The method includes: obtaining output data of a first-class photovoltaic power station in real time, obtaining output data of a second-class photovoltaic power station in a specified area at specified intervals, and restoring the missing output data of the second-class photovoltaic power station in the specified area; analyzing the output data of the first-class photovoltaic power station and the second-class photovoltaic power station to select photovoltaic power stations that meet the requirements; performing cluster analysis on the photovoltaic power stations that meet the requirements, and selecting the photovoltaic power station closest to the cluster center as a model station; calculating the numerical weather forecast results of the model station, inputting the numerical weather forecast results of the model station into a preset photovoltaic power generation model to obtain the predicted power of the model station; using statistical upscaling to calculate the predicted power of each model station to obtain a predicted value of the power of the entire network; and formulating a control strategy corresponding to the predicted value of the power of the entire network. This invention improves the accuracy of predicting distributed photovoltaic power generation, thereby improving the quality of control.

[0006] The above-mentioned existing technologies all have the problems raised in the background technology of this application: they do not fully consider the coordinated optimization between energy storage equipment, photovoltaic power generation and user electricity consumption, and are unable to maximize the advantages of energy storage equipment and improve energy utilization efficiency. Summary of the Invention

[0007] The present invention is intended to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Simplifications or omissions may be made in the present invention and the abstract and title of this application to avoid obscuring the purpose of the present invention, the abstract and the title of the application, and such simplifications or omissions shall not be used to limit the scope of the present invention.

[0008] In order to solve the above technical problems, the present invention provides a power control system and device based on photovoltaic intelligent micro-breaker.

[0009] In one aspect, the present invention provides a power control device based on a photovoltaic intelligent micro-breaker, comprising a data acquisition unit, a virtual power storage unit, a strategy unit, and a control instruction unit; wherein: The data acquisition unit is used to obtain grid operation parameters and collect energy storage equipment data, power generation data, and load data of distributed photovoltaic users based on photovoltaic intelligent micro-breakers; The strategy unit is equipped with a deep reinforcement learning model. In each control cycle, it generates control actions for adjusting the inverter output power of each photovoltaic user and the charging power and discharging power of the energy storage device based on the power generation data, load data, energy storage device data, and grid operating parameters. It then selects the optimal control action to generate the power generation strategy and energy storage strategy for each distributed photovoltaic user. The virtual power storage unit is used to record and update the energy storage device data, power generation data, load data, and grid operation parameters; the virtual power storage unit is also used to simulate the execution of each control action during the training process of the deep reinforcement learning model and calculate the expected return of each control action; The control instruction unit generates a control instruction for each photovoltaic user based on the power generation strategy and the energy storage strategy, and sends each control instruction to the corresponding photovoltaic intelligent micro-breaker.

[0010] As a preferred solution of the power control device based on a photovoltaic intelligent micro-breaker of the present invention, the data acquisition unit includes a micro-breaker sub-unit and a grid sub-unit; wherein the micro-breaker sub-unit is used to communicate with the photovoltaic intelligent micro-breaker of each distributed photovoltaic user to collect energy storage device data, power generation data, and load data of each photovoltaic user; The energy storage device data includes the current power, maximum capacity, minimum capacity, rated charging power, and rated discharge power of the energy storage device; the power generation data includes the output power and current maximum output power of the photovoltaic inverter; the load data is the power load power of the photovoltaic user; the grid subunit is used to obtain grid operation parameters; the grid operation parameters include grid load and real-time electricity price.

[0011] As a preferred solution of the power control device based on a photovoltaic intelligent micro-breaker of the present invention, the strategy unit includes a strategy model subunit; the strategy model subunit is configured with a deep reinforcement learning model for outputting the selection probability of each control action; The deep reinforcement learning model includes a state space, an action space, and a policy network; wherein the state space includes the energy storage device data, power generation data, load data, and feature vectors of grid operation parameters of all photovoltaic users; The action space includes the eigenvectors of all control actions; wherein any control action includes the charging power or discharging power of the energy storage device of each photovoltaic user and the output power adjustment of each photovoltaic inverter in the current regulation cycle; The input of the policy network includes all feature vectors contained in the state space and the action space; the output includes the selection probability of each control action in the action space.

[0012] As a preferred solution of the power control device based on a photovoltaic intelligent micro-breaker of the present invention, the strategy unit further includes a data sorting subunit; the data sorting subunit is used to sort the energy storage device data, power generation data, load data and grid operation parameters to generate the state space and action space; the details are as follows: Each item of energy storage device data, power generation data, load data, and grid operation parameters is cleaned, normalized, and encoded to obtain each eigenvector contained in the state space; A charging power is set for each energy storage device based on the rated charging power, or a discharge power is set for each energy storage device based on the rated discharge power, and an output power adjustment amount is set for each photovoltaic inverter based on the output power and the current maximum output power of each photovoltaic inverter, forming a set of control parameters; M groups of control parameters are repeatedly generated, and the M groups of control parameters are respectively normalized and encoded to obtain feature vectors of M control actions and form an action space.

[0013] As a preferred solution of the power control device based on a photovoltaic intelligent micro-breaker of the present invention, the data sorting subunit is further used to select the optimal control action and generate the power generation strategy and energy storage strategy for each distributed photovoltaic user based on the optimal control action; the details are as follows: Count the selection probability of each control action and extract the control action with the highest selection probability as the best control action; Decoding the optimal control action to obtain the charging power or discharging power of the energy storage device of each photovoltaic user included in the optimal control action as the energy storage strategy of the corresponding photovoltaic user; The output power adjustment amount of each photovoltaic inverter included in the optimal control action is obtained and used as the power generation strategy of the corresponding photovoltaic user.

[0014] As a preferred solution of the power control device based on a photovoltaic intelligent micro-breaker of the present invention, the virtual power storage unit includes a virtual power storage station; the virtual power storage station is used to record and update energy storage device data, power generation data, load data, and grid operation parameters; the virtual power storage station is also used to simulate the execution of each control action during the training process of the policy network; the details are as follows: The virtual energy storage station updates the output power of each photovoltaic inverter based on the output power adjustment of each photovoltaic inverter; and calculates and updates the current power of each energy storage device based on the charging power or discharging power of each energy storage device.

[0015] As a preferred solution of the power control device based on a photovoltaic intelligent micro-breaker of the present invention, the virtual power storage unit further includes a calculation subunit; the calculation subunit is configured with a reward function; the calculation subunit calculates the expected reward of each control action based on the simulated execution of each control action; the formula of the reward function is as follows: ; Where R represents the expected reward of executing any control action; Indicates the scheduling reward value for executing the corresponding control action; represents the benefit reward value of the i-th PV user after executing the corresponding control action; the value range of i is 1, 2, ..., n; n is the number of PV users; Represents the loss penalty value of the i-th photovoltaic user after executing the corresponding control action; if after executing the corresponding control action, the current power of the energy storage device corresponding to the i-th photovoltaic user is less than the corresponding minimum capacity or greater than the corresponding maximum capacity, then The value is 1, otherwise, The value is 0; Indicates the power penalty factor for executing the corresponding control action; calculates the total power uploaded to the grid by all photovoltaic users after executing the control action; the calculation subunit is configured with a total power threshold. If the total power is greater than the total power threshold, then is the ratio of the total power to the total power threshold, otherwise, is 0; are all weight coefficients.

[0016] As a preferred solution of the power control device based on photovoltaic intelligent micro-breaker of the present invention, the calculation formula of the scheduling reward value is as follows: ; in, represents the charging power of the energy storage device of the i-th PV user after the control action is executed; represents the discharge power of the energy storage device of the i-th photovoltaic user after the control action is executed; L represents the grid load before the control action is executed; Indicates the peak load threshold of the power grid preset in the calculation subunit; Indicates the valley load threshold of the power grid preset in the calculation subunit; Both are adjustment coefficients.

[0017] As a preferred solution of the power control device based on photovoltaic intelligent micro-breaker of the present invention, the calculation formula of the benefit reward value of the photovoltaic user is as follows: ; in, represents the output power of the PV inverter of the i-th PV user after the control action is executed; represents the power load power of the i-th photovoltaic user; T represents the length of the control cycle; e represents the real-time electricity price; The electricity price threshold value pre-set for the calculation subunit.

[0018] On the other hand, the present invention provides a power control system based on a photovoltaic intelligent micro-breaker, comprising an intelligent micro-breaker module and a power control device based on a photovoltaic intelligent micro-breaker according to the present invention; wherein: The intelligent micro-break module includes a photovoltaic intelligent micro-breaker installed at each distributed photovoltaic user; the intelligent micro-break module is used to collect energy storage equipment data, power generation data, and load data of distributed photovoltaic users and transmit them to the power control device; The power control device collects grid operating parameters, and combines the energy storage device data, power generation data, and load data of the distributed photovoltaic users to generate a power generation strategy and energy storage strategy for each photovoltaic user; the power control device generates a control instruction for each photovoltaic user based on the power generation strategy and energy storage strategy, and sends each control instruction to the corresponding photovoltaic intelligent micro-breaker; the photovoltaic intelligent micro-breaker parses the control instruction, obtains the power generation strategy and energy storage strategy, and controls the output power of the photovoltaic inverter of the corresponding photovoltaic user based on the power generation strategy, and controls the charging power or discharging power of the energy storage device of the corresponding photovoltaic user based on the energy storage strategy.

[0019] The beneficial effects of the present invention are as follows: By generating control actions through the strategy unit, the system effectively adjusts the inverter output power of photovoltaic users and the charging and discharging power of energy storage devices, achieving peak load shifting and valley shifting for the grid, reducing the risk of excessive total power uploaded to the grid by distributed photovoltaic users, and ensuring safe and stable grid operation. A deep reinforcement learning model is configured to automatically generate and select optimal control actions based on comprehensive data, forming power generation and energy storage strategies. Compared to traditional manual or simple preset strategies, this approach can better adapt to complex and changing grid environments and user electricity demands, achieving intelligent dynamic power regulation.

[0020] A reward function is used to measure the impact of control actions on users. Based on real-time electricity prices and user load, photovoltaic users can better achieve economic benefits through peak load shifting. Factors such as the energy storage device's power and capacity are considered to avoid overcharging and discharging, extending the device's service life. Energy storage devices are also rationally utilized to coordinate photovoltaic power generation with user electricity demand, improving energy efficiency. This solves the problem of household distributed energy storage devices being unable to fully utilize their regulatory role in the power grid due to their decentralized nature, small capacity, and lack of effective coordination mechanisms, thereby enhancing the support capacity of energy storage resources for the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be derived from these drawings without inventive effort. Among them: Figure 1 A schematic diagram of the structure of a power control device based on a photovoltaic intelligent micro-breaker provided by the present invention; Figure 2Schematic diagram of the data interaction mode between the photovoltaic intelligent micro-breaker and the power control device provided by the present invention. DETAILED DESCRIPTION

[0022] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0023] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it constitute a separate or selective embodiment that is mutually exclusive with other embodiments.

[0025] Example 1 like Figure 1 As shown, a power control device based on a photovoltaic intelligent micro-breaker includes a data acquisition unit, a virtual power storage unit, a strategy unit, and a control instruction unit; wherein: The data acquisition unit is used to obtain grid operation parameters and collect energy storage equipment data, power generation data, and load data of distributed photovoltaic users based on photovoltaic intelligent micro-breakers; The data acquisition unit includes a micro-break sub-unit and a grid sub-unit; wherein the micro-break sub-unit is used to communicate with the photovoltaic intelligent micro-breaker of each distributed photovoltaic user to collect energy storage equipment data, power generation data, and load data of each photovoltaic user; The energy storage device data includes the current power, maximum capacity, minimum capacity, rated charging power, and rated discharging power of the energy storage device; The maximum capacity of an energy storage device is its upper threshold capacity for charging, and the minimum capacity is its lower threshold capacity for discharging. When charging and discharging the energy storage device, its power needs to be controlled between the upper threshold capacity and the lower threshold capacity to avoid overcharging or over-discharging causing additional damage to the energy storage device.

[0026] The power generation data includes the output power and current maximum output power of the photovoltaic inverter. The output power is the actual output power of the photovoltaic power generation, while the current maximum output power is the maximum power the inverter can output under current lighting conditions. Inverters often limit their output power due to factors such as grid saturation, fully charged energy storage devices, and low user load. The inverter can obtain key parameters such as the current DC input voltage and current and calculate the current maximum output power based on the inverter's power-voltage curve or power-current curve. Furthermore, the impact of the inverter's internal component temperature on power output can be considered. Inverter manufacturers generally provide power derating curves for different temperature conditions. Based on the current inverter temperature monitoring value, the corresponding power correction factor is found in the derating curve to correct for a more accurate maximum output power. If the calculated current maximum output power is greater than the rated output power of the photovoltaic inverter, the current maximum output power value is recorded as the rated output power of the photovoltaic inverter.

[0027] The load data is the electricity load power of photovoltaic users.

[0028] The grid subunit is used to obtain grid operation parameters; the grid operation parameters include grid load and real-time electricity price.

[0029] The strategy unit is equipped with a deep reinforcement learning model. In each control cycle, it generates control actions for adjusting the inverter output power of each photovoltaic user and the charging power and discharging power of the energy storage device based on the energy storage device data, power generation data, load data, and grid operating parameters. It then selects the optimal control action to generate the power generation strategy and energy storage strategy for each distributed photovoltaic user. The strategy unit includes a strategy model subunit and a data sorting subunit; wherein the strategy model subunit is configured with a deep reinforcement learning model to output the selection probability of each control action; The deep reinforcement learning model includes a state space, an action space, and a policy network; wherein the state space includes the energy storage device data, power generation data, load data, and feature vectors of grid operation parameters of all photovoltaic users; The action space includes the feature vectors of all optional control actions; wherein any control action includes the charging power or discharging power of the energy storage device of each photovoltaic user and the output power adjustment of each photovoltaic inverter in the current regulation cycle; The input of the policy network includes all feature vectors contained in the state space and the action space; the output includes the selection probability of each optional control action in the action space.

[0030] The training method of the policy network is as follows: S100, constructing a policy network and initializing its parameters; the data sorting subunit simulates and generates state space and action space; setting the maximum number of iterations; The policy network is a deep neural network, including an input layer, a hidden layer, and an output layer; wherein the input layer is used to input feature vectors of the state space and action space; the hidden layer is used to extract the features of the input feature vectors; and the output layer is used to map the features of the input feature vectors to the selection probability of each control action.

[0031] The training of the policy network requires a large amount of energy storage equipment data, power generation data, load data and grid operating parameters; numerical simulation is used to generate a state space with a relatively comprehensive value range, and the action space is generated based on the state space simulation. These simulated data are used as the data source for training the policy network.

[0032] S200, inputting the feature vectors of the state space and the action space into the policy network, and selecting a control action based on the output of the policy network; Control actions can be selected using either a greedy or ε-greedy strategy. The policy network outputs the probability of each control action. The greedy strategy selects the control action with the highest probability. The ε-greedy strategy selects the control action with the highest probability with a probability of 1-ε and randomly selects a control action with a probability of ε. The ε-greedy strategy introduces exploration of other control actions, helping to discover valuable control actions that might otherwise be overlooked.

[0033] S300, simulating the execution of the selected control action through the virtual power storage unit, updating the state space and action space, and calculating the expected return of executing the control action; After executing a control action, the energy storage device data for the next control cycle will change. The virtual energy storage unit calculates the value of the energy storage device data that changed after the control action was executed at the beginning of the next control cycle, updates the state space accordingly, and then updates the action space based on the state space to prepare for the simulation of the next control cycle. In addition, the virtual energy storage unit can also draw change curves for power generation data and load data based on the user's historical power generation data and load data, simulate changes in power generation data and load data based on these change curves, and update the state space, thereby providing more comprehensive data for policy network training.

[0034] S400, calculating the cumulative reward value of each iterative execution of the control action based on the expected reward, and updating the parameters of the policy network using a gradient descent method based on the cumulative reward value; The formula for calculating the cumulative return value is as follows: ; in, Represents the current cumulative reward value; N represents the number of control actions that have been simulated and executed, that is, the current number of iterations; Represents the discount factor, the value range is (0, 1], Represents the discount factor the jth power of represents the expected return of executing the control action at the jth iteration; the value range of j is 1, 2, ..., N; The calculation formula for updating the parameters of the policy network is as follows: ; in, represents any parameter in the policy network; Indicates the function pair in brackets Find the gradient; is the learning rate; is the loss function, and the calculation formula is as follows: ; in, represents the selection probability of the control action executed in the jth iteration, that is, the selection probability assigned by the policy network to the control action.

[0035] After the above parameter updates, the policy network will assign a greater probability of selecting the control action that increases the cumulative reward value; this update process will be repeated continuously, and the policy network will gradually learn the optimal probability distribution of each control action in different state spaces to maximize the cumulative reward value.

[0036] S500, repeating steps S200-S400 until the maximum number of iterations is reached, completing the training of the policy network; saving the trained policy network and deploying the application.

[0037] After multiple iterations, the policy network is able to make decisions that maximize the cumulative reward value, that is, the policy network can assign the maximum selection probability to those control actions that maximize the cumulative reward value, thereby guiding the selection of the best control action.

[0038] The data collating subunit is used to collate the energy storage device data, power generation data, load data, and grid operation parameters to generate the state space and action space; specifically, as follows: Each item of energy storage device data, power generation data, load data, and grid operation parameters is cleaned, normalized, and encoded to obtain each eigenvector contained in the state space; A charging power is set for each energy storage device based on the rated charging power, or a discharge power is set for each energy storage device based on the rated discharge power, and the output power adjustment amount is set for each photovoltaic inverter based on the output power and the current maximum output power of each photovoltaic inverter, forming a set of control parameters; M groups of control parameters are repeatedly generated, and the M groups of control parameters are normalized and encoded respectively to obtain feature vectors of M control actions and form an action space; M is a positive integer.

[0039] For example, each energy storage device can be controlled to charge or discharge, with the discharge power set to a random value between 0 and the rated discharge power; or the charge power set to a random value between 0 and the rated charge power. The output power of the PV inverter can be increased or decreased based on the inverter's output power and the current maximum output power, for example, by setting the output power adjustment amount of the PV inverter with an adjustment step of 5% or 10% of the current maximum output power. The set control parameters are normalized and then encoded, for example, using binary amplitude encoding. For charge and discharge power, using a control step of 20% of the rated power, 0001 indicates a charge power of 20% of the rated charge power, and 1010 indicates a discharge power of 40% of the rated discharge power. The first bit of the encoding is 0 for charging and 1 for discharging. For the inverter output power, using a control step of 5% of the current maximum output power, for example, 0001 indicates an increase in the inverter output power of 5%, and 1010 indicates a decrease in the inverter output power of 10%. The first bit of the encoding is 0 for an increase in the inverter output power, and 1 for a decrease in the inverter output power. When the grid load is low and household electricity load is met, the action selected by the strategic network may be to reduce the inverter power output. When the grid needs more power support and photovoltaic power generation conditions permit, the action selected may be to increase the inverter power output.

[0040] The data sorting subunit is further configured to select an optimal control action and generate a power generation strategy and energy storage strategy for each distributed photovoltaic user based on the optimal control action; specifically, as follows: Count the selection probability of each control action and extract the control action with the highest selection probability as the best control action; Decoding the optimal control action to obtain the charging power or discharging power of the energy storage device of each photovoltaic user included in the optimal control action as the energy storage strategy of the corresponding photovoltaic user; The output power adjustment amount of each photovoltaic inverter included in the optimal control action is obtained and used as the power generation strategy of the corresponding photovoltaic user.

[0041] The virtual power storage unit is used to record and update the energy storage device data, power generation data, load data and grid operation parameters; the virtual power storage unit is also used to simulate the execution of each control action during the training process of the strategy network and calculate the expected return of each control action; The virtual power storage unit includes a computing subunit and a virtual power storage station; wherein the virtual power storage station is used to record and update energy storage equipment data, power generation data, load data and grid operation parameters; specifically as follows: Data on energy storage devices, power generation data, load data, and grid operating parameters are cleaned and standardized. Massive amounts of collected data are cleaned to remove erroneous or abnormal data points. For example, data outside the normal operating range of the equipment is corrected or deleted. The data is then standardized, converting different types of data into a unified dimension and format.

[0042] Each PV user's energy storage device data, power generation data, and load data are stored on a cloud computing platform, forming a virtual storage power station. Grid operating parameters are then added to the virtual storage power station. The cleansed and standardized data is then stored using the cloud computing platform's storage services. Cloud computing platforms can provide highly reliable and scalable data storage solutions that meet the long-term storage needs of massive amounts of data. For example, distributed file systems (such as HDFS) or cloud databases (such as Amazon DynamoDB) can be used to store data. These storage systems automatically scale storage capacity based on data growth and provide efficient data read and write interfaces.

[0043] A data update cycle is set, and at the beginning of each data update cycle, all data items in the virtual energy storage power station are collected and updated. Based on the data update cycle, the data acquisition unit regularly collects energy storage device data, power generation data, load data storage, and grid operating parameters for each photovoltaic user and uploads them to the cloud computing platform, ensuring that the virtual energy storage power station can promptly reflect the overall energy storage status and grid conditions.

[0044] A storage station is a common form of large-scale centralized energy storage, usually consisting of a large number of energy storage battery packs, power conversion systems, monitoring systems, etc. It can store a large amount of electrical energy and discharge it when the grid needs it, playing the role of peak shaving and valley filling, frequency regulation and voltage regulation, etc. In this application, the various distributed energy storage devices are not concentrated in a single storage station, but through the unified regulation of the apparatus and system described in this application, the distributed energy storage devices can be centralized to complete the peak shaving and valley filling of the grid load, achieving the function equivalent to a storage station. The data of each energy storage device is stored, monitored, updated and simulated using the same cloud computing platform. The management mode of this energy storage device is also similar to that of a real storage station.

[0045] The virtual power station is also used to simulate and execute each control action during the training process of the policy network; specifically as follows: The virtual storage power station updates the output power of each PV inverter based on its output power adjustment. It also calculates and updates the current capacity of each energy storage device based on its charging or discharging power. The increase in the energy storage device's capacity is calculated by multiplying the charging power by the length of the control cycle, while the decrease in the energy storage device's capacity is calculated by multiplying the discharge power by the length of the control cycle. The increase is added to the current capacity of the energy storage device, or the decrease is subtracted from the current capacity to obtain the updated value, which is the energy storage device's capacity at the start of the next control cycle.

[0046] The calculation subunit is configured with a reward function; the calculation subunit calculates the expected reward of each control action based on the simulated execution of each control action; the formula of the reward function is as follows: ; Where R represents the expected reward of executing any control action; Indicates the scheduling reward value for executing the corresponding control action; represents the benefit reward value of the i-th PV user after executing the corresponding control action; the value range of i is 1, 2, ..., n; n is the number of PV users; Represents the loss penalty value of the i-th photovoltaic user after executing the corresponding control action; if after executing the corresponding control action, the current power of the energy storage device corresponding to the i-th photovoltaic user is less than the corresponding minimum capacity or greater than the corresponding maximum capacity, then The value is 1, otherwise, The value is 0; Indicates the power penalty factor for executing the corresponding control action; calculates the total power uploaded to the grid by all photovoltaic users after executing the control action; the calculation subunit is configured with a total power threshold. If the total power is greater than the total power threshold, then is the ratio of the total power to the total power threshold, otherwise, is 0; are all weight coefficients, which are set by those skilled in the art based on actual needs; in the formula of the reward function, each parameter is normalized and dimensionless.

[0047] The calculation formula of the scheduling reward value is as follows: ; in, represents the charging power of the energy storage device of the i-th PV user after the control action is executed; represents the discharge power of the energy storage device of the i-th photovoltaic user after the control action is executed; L represents the grid load before the control action is executed; Indicates the peak load threshold of the power grid preset in the calculation subunit; Indicates the valley load threshold of the power grid preset in the calculation subunit; These are adjustment coefficients and are set by those skilled in the art based on actual needs.

[0048] The dispatch reward value is used to measure the positive impact of executing the control action on the peak load reduction and valley filling of the power grid. When the grid load L is greater than the peak load threshold, it is considered that the power grid is at peak load at this time, and the cumulative discharge power of distributed photovoltaic users can alleviate the peak load pressure of the power grid; and the dispatch reward value is weighted based on the grid load. The higher the grid load, the more urgent the need for peak load reduction, and the greater the dispatch reward value. During peak hours of electricity consumption, such as in the afternoons in summer, the grid load will rise sharply due to the use of a large number of high-power electrical appliances such as air conditioners. If peak load reduction is not performed, the power generation equipment may not be able to meet all electricity demand. Peak load reduction through charging of energy storage equipment can make the power demand curve flatter. In this way, the power generation side can supply electricity more stably, avoiding problems such as grid voltage drop and frequency fluctuation caused by sudden demand exceeding supply capacity.

[0049] When the grid load L is less than the valley load threshold, the grid is considered to be at valley load. The cumulative charging power of distributed photovoltaic users can stabilize the grid's valley load. The dispatch reward value is weighted based on the grid load. The lower the grid load, the more urgent the need to fill the valley, and the larger the dispatch reward value. During low electricity consumption periods, such as late at night, when electricity demand is low, many power generation equipment may be operating at low power or even shut down. Charging energy storage devices to fill the valley can improve the utilization rate of power generation equipment, allowing it to operate more stably within a reasonable power range, avoid frequent starts and stops, and reduce the impact on the grid during equipment startup and shutdown. At the same time, it can better absorb new energy and reduce the phenomenon of wind and solar power curtailment.

[0050] The calculation formula of the benefit reward value of the photovoltaic user is as follows: ; in, represents the output power of the PV inverter of the i-th PV user after the control action is executed; represents the power load power of the i-th photovoltaic user; T represents the length of the control cycle; e represents the real-time electricity price; The electricity price threshold value pre-set for the calculation subunit.

[0051] The benefit reward value is used to measure the economic benefits obtained by each photovoltaic user through peak load shifting. , at this time, it is low-priced electricity, encouraging users to purchase more electricity from the grid; when the real-time electricity price e is greater than or equal to the electricity price threshold At this time, the price of electricity is high, which encourages users to sell more electricity to the grid; The calculation formula for the total power uploaded to the grid by all photovoltaic users is as follows: ; in, This value represents the total power uploaded to the grid by all PV users. If the total power uploaded to the grid by distributed PV users is too high, the voltage connected to the grid may be too high, even exceeding the rated range of the grid transformer and damaging it. Therefore, it is necessary to limit the total power uploaded to the grid by PV users.

[0052] When calculating the expected return, if any energy storage device is discharged, its charging power is the discharge power multiplied by -1; conversely, when calculating the cumulative discharge power, if any energy storage device is charged, its charging power is the charging power multiplied by -1.

[0053] The control instruction unit generates a control instruction for each photovoltaic user based on the power generation strategy and the energy storage strategy, and sends each control instruction to the corresponding photovoltaic intelligent micro-breaker.

[0054] Example 2 Based on the same inventive concept as Example 1, this embodiment introduces a power control system based on a photovoltaic intelligent micro-breaker, including an intelligent micro-breaker module and a power control device based on a photovoltaic intelligent micro-breaker as described in Example 1; wherein the intelligent micro-breaker module includes a photovoltaic intelligent micro-breaker installed at each distributed photovoltaic user; the intelligent micro-breaker module is used to collect energy storage equipment data, power generation data, and load data of the distributed photovoltaic user, and transmit them to the power control device; the power control device collects grid operating parameters, and combines the energy storage equipment data, power generation data, and load data of the distributed photovoltaic user to generate a power generation strategy and energy storage strategy for each distributed photovoltaic user; the power control device generates a control instruction for each photovoltaic user based on the power generation strategy and energy storage strategy, and sends each control instruction to the corresponding photovoltaic intelligent micro-breaker; the photovoltaic intelligent micro-breaker parses the control instruction, obtains the power generation strategy and energy storage strategy, and controls the output power of the photovoltaic inverter of the corresponding photovoltaic user based on the power generation strategy, and controls the charging power or discharging power of the energy storage device of the corresponding photovoltaic user based on the energy storage strategy.

[0055] PV smart MCBs feature an expandable power parameter interface that allows them to communicate with energy storage devices to obtain their current power level, maximum capacity, minimum capacity, rated charging power, and rated discharge power. They can also communicate with PV inverters to read their output power and current maximum output power. Furthermore, many smart meters now include communication capabilities, such as power line carrier communication and wireless communication, to exchange data with external devices. By adapting these communication protocols, smart MCBs can connect to household main meters and read PV users' electricity load power. Grid information such as real-time electricity prices and grid load power can be obtained by power control devices through data interfaces with grid operators or electricity market data platforms.

[0056] The data interaction mode between photovoltaic intelligent micro-breaker and power control device is as follows Figure 2 For example, a photovoltaic smart micro-breaker can send instructions to the photovoltaic inverter to adjust its output power and operating mode. It can also send instructions to the energy storage device to control its charging and discharging, coordinating it with photovoltaic power generation to meet user electricity needs or supply power to the grid. By precisely controlling the discharge power of the energy storage device and the photovoltaic power generation, the energy storage device's role in grid peak regulation is maximized while ensuring user power reliability. This can further improve the energy efficiency and stability of distributed photovoltaic systems and achieve efficient coordinated operation between photovoltaic power generation, energy storage, and the grid.

[0057] It is important to note that the configuration and arrangement of the present application, as illustrated in various exemplary embodiments, are illustrative only. Although only two embodiments are described in detail in this disclosure, those reading this disclosure will readily appreciate that numerous modifications are possible without materially departing from the novel teachings and advantages of the subject matter described herein, including, for example, variations in the size, dimensions, structure, shape, and proportions of various components, as well as parameter values ​​(e.g., temperature, pressure, etc.), mounting arrangements, use of materials, color, orientation, and the like. For example, components shown as integrally formed may be comprised of multiple parts or components, the positions of components may be inverted or otherwise altered, and the nature, number, or position of discrete components may be modified or changed. Therefore, all such modifications are intended to be encompassed within the scope of this invention. The order or sequence of any process or method steps may be altered or reordered according to alternative embodiments. Any "means-plus-function" clause is intended to cover structures that perform the functions described herein, and not only structural equivalence but also structural equivalents. Other substitutions, modifications, changes, and omissions may be made in the design, operating conditions, and arrangement of the exemplary embodiments without departing from the scope of this invention. Therefore, the invention is not limited to the specific embodiments, but extends to various modifications that still fall within the scope of the appended claims.

[0058] Additionally, in order to provide a concise description of exemplary embodiments, all features of an actual embodiment may not be described (i.e., those features that are not relevant to the best mode presently contemplated for carrying out the invention or those that are not relevant to implementing the invention).

[0059] It should be understood that in the development of any actual embodiment, as in any engineering or design project, numerous implementation-specific decisions may be made. Such a development effort may be complex and time-consuming, but for those of ordinary skill having the benefit of this disclosure, the development effort will be a routine task of design, fabrication, and production without undue experimentation.

[0060] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A power control device based on a photovoltaic intelligent micro-breaker, characterized by: It includes data acquisition unit, virtual power storage unit, strategy unit and control instruction unit; among which: The data acquisition unit is used to obtain grid operation parameters and collect energy storage equipment data, power generation data, and load data of distributed photovoltaic users based on photovoltaic intelligent micro-breakers; The strategy unit is equipped with a deep reinforcement learning model. In each control cycle, it generates control actions for adjusting the inverter output power of each photovoltaic user and the charging power and discharging power of the energy storage device based on the energy storage device data, power generation data, load data, and grid operating parameters. It then selects the optimal control action to generate the power generation strategy and energy storage strategy for each distributed photovoltaic user. The virtual power storage unit is used to record and update the energy storage device data, power generation data, load data, and grid operation parameters; the virtual power storage unit is also used to simulate the execution of each control action during the training process of the deep reinforcement learning model and calculate the expected return of each control action; The control instruction unit generates a control instruction for each photovoltaic user based on the power generation strategy and the energy storage strategy, and sends each control instruction to the corresponding photovoltaic intelligent micro-breaker.

2. The power control device based on a photovoltaic intelligent micro-breaker according to claim 1, characterized in that: The data acquisition unit includes a micro-break sub-unit and a grid sub-unit; wherein the micro-break sub-unit is used to communicate with the photovoltaic intelligent micro-breaker of each distributed photovoltaic user to collect energy storage equipment data, power generation data, and load data of each photovoltaic user; The energy storage device data includes the current power, maximum capacity, minimum capacity, rated charging power, and rated discharge power of the energy storage device; the power generation data includes the output power and current maximum output power of the photovoltaic inverter; the load data is the power load power of the photovoltaic user; the grid subunit is used to obtain grid operation parameters; the grid operation parameters include grid load and real-time electricity price.

3. The power control device based on a photovoltaic intelligent micro-breaker according to claim 2, characterized in that: The strategy unit includes a strategy model subunit; the strategy model subunit is configured with a deep reinforcement learning model to output the selection probability of each control action; The deep reinforcement learning model includes a state space, an action space, and a policy network; wherein the state space includes the energy storage device data, power generation data, load data, and feature vectors of grid operation parameters of all photovoltaic users; The action space includes the eigenvectors of all control actions; wherein any control action includes the charging power or discharging power of the energy storage device of each photovoltaic user and the output power adjustment of each photovoltaic inverter in the current regulation cycle; The input of the policy network includes all feature vectors contained in the state space and the action space; the output includes the selection probability of each control action in the action space.

4. The power control device based on a photovoltaic intelligent micro-breaker according to claim 3, characterized in that: The strategy unit also includes a data sorting subunit; the data sorting subunit is used to sort the energy storage device data, power generation data, load data and grid operation parameters to generate the state space and action space; specifically as follows: Each item of energy storage device data, power generation data, load data, and grid operation parameters is cleaned, normalized, and encoded to obtain each eigenvector contained in the state space; A charging power is set for each energy storage device based on the rated charging power, or a discharge power is set for each energy storage device based on the rated discharge power, and an output power adjustment amount is set for each photovoltaic inverter based on the output power and the current maximum output power of each photovoltaic inverter, forming a set of control parameters; M groups of control parameters are repeatedly generated, and the M groups of control parameters are respectively normalized and encoded to obtain feature vectors of M control actions and form an action space.

5. The power control device based on a photovoltaic intelligent micro-breaker according to claim 4, characterized in that: The data sorting subunit is further configured to select an optimal control action and generate a power generation strategy and energy storage strategy for each distributed photovoltaic user based on the optimal control action; specifically, as follows: Count the selection probability of each control action and extract the control action with the highest selection probability as the best control action; Decoding the optimal control action to obtain the charging power or discharging power of the energy storage device of each photovoltaic user included in the optimal control action as the energy storage strategy of the corresponding photovoltaic user; The output power adjustment amount of each photovoltaic inverter included in the optimal control action is obtained and used as the power generation strategy of the corresponding photovoltaic user.

6. The power control device based on a photovoltaic intelligent micro-breaker according to claim 5, characterized in that: The virtual power storage unit includes a virtual power station; the virtual power station is used to record and update energy storage equipment data, power generation data, load data and grid operation parameters; the virtual power station is also used to simulate the execution of each control action during the training process of the strategy network; the details are as follows: The virtual energy storage station updates the output power of each photovoltaic inverter based on the output power adjustment of each photovoltaic inverter; and calculates and updates the current power of each energy storage device based on the charging power or discharging power of each energy storage device.

7. The power control device based on a photovoltaic intelligent micro-breaker according to claim 6, characterized in that: The virtual power storage unit further includes a calculation subunit; the calculation subunit is configured with a reward function; the calculation subunit calculates the expected reward of each control action based on the simulated execution of each control action; the formula of the reward function is as follows: ； Where R represents the expected reward of executing any control action; Indicates the scheduling reward value for executing the corresponding control action; represents the benefit reward value of the i-th PV user after executing the corresponding control action; the value range of i is 1, 2, ..., n; n is the number of PV users; Represents the loss penalty value of the i-th photovoltaic user after executing the corresponding control action; if after executing the corresponding control action, the current power of the energy storage device corresponding to the i-th photovoltaic user is less than the corresponding minimum capacity or greater than the corresponding maximum capacity, then The value is 1, otherwise, The value is 0; Indicates the power penalty factor for executing the corresponding control action; calculates the total power uploaded to the grid by all photovoltaic users after executing the control action; the calculation subunit is configured with a total power threshold. If the total power is greater than the total power threshold, then is the ratio of the total power to the total power threshold, otherwise, is 0; are all weight coefficients.

8. The power control device based on a photovoltaic intelligent micro-breaker according to claim 7, characterized in that: The calculation formula of the scheduling reward value is as follows: ； in, represents the charging power of the energy storage device of the i-th PV user after the control action is executed; represents the discharge power of the energy storage device of the i-th photovoltaic user after the control action is executed; L represents the grid load before the control action is executed; Indicates the peak load threshold of the power grid preset in the calculation subunit; Indicates the valley load threshold of the power grid preset in the calculation subunit; Both are adjustment coefficients.

9. The power control device based on a photovoltaic intelligent micro-breaker according to claim 8, characterized in that: The calculation formula of the benefit reward value of the photovoltaic user is as follows: ； in, represents the output power of the PV inverter of the i-th PV user after the control action is executed; represents the power load power of the i-th photovoltaic user; T represents the length of the control cycle; e represents the real-time electricity price; The electricity price threshold value pre-set for the calculation subunit.

10. A power control system based on photovoltaic intelligent micro-breaker, characterized by: It includes an intelligent micro-breaker module and a power control device based on a photovoltaic intelligent micro-breaker according to any one of claims 1 to 9; wherein: The intelligent micro-break module includes a photovoltaic intelligent micro-breaker installed at each distributed photovoltaic user; the intelligent micro-break module is used to collect energy storage equipment data, power generation data, and load data of distributed photovoltaic users and transmit them to the power control device; The power control device collects grid operating parameters, and combines the energy storage device data, power generation data, and load data of the distributed photovoltaic users to generate a power generation strategy and energy storage strategy for each photovoltaic user; the power control device generates a control instruction for each photovoltaic user based on the power generation strategy and energy storage strategy, and sends each control instruction to the corresponding photovoltaic intelligent micro-breaker; the photovoltaic intelligent micro-breaker parses the control instruction, obtains the power generation strategy and energy storage strategy, and controls the output power of the photovoltaic inverter of the corresponding photovoltaic user based on the power generation strategy, and controls the charging power or discharging power of the energy storage device of the corresponding photovoltaic user based on the energy storage strategy.

Images