Micro-grid intelligent control method and device based on multi-agent cooperation

Abstract

The invention discloses a micro-grid intelligent control method and a micro-grid intelligent control device based on multi-agent cooperation, and aims to solve the problems that the existing micro-grid intelligent control method uses limited data, and new energy generating capacity prediction information, current environment information, load information and the like are difficult to put into a model for decision making. The invention provides a method for constructing a micro-grid control large model, and aims to solve the problems that a multi-energy electric quantity prediction agent, a multi-load prediction agent and an energy storage agent are constructed on the basis of a large model base, and a control strategy of large and small models and multi-agent cooperation is realized. According to the method, a more comprehensive decision-making system is constructed, so that more intelligent and precise strategy control of the micro-grid is realized, and the economic benefit of multi-energy complementation and cooperation is improved.

Description

Technical Field

[0001] The present invention belongs to the field of microgrids, and in particular relates to a microgrid intelligent control method and device based on multi-agent collaboration. Background Art

[0002] With the continuous growth of energy demand and concerns about the stability and environmental protection of traditional energy supply, microgrids as a distributed energy system have received widespread attention. At present, microgrid control faces challenges. For example, in a complex operating environment, traditional control methods are difficult to achieve efficient coordinated control of multiple distributed energy resources (such as solar panels, wind turbines, energy storage equipment, etc.), resulting in low energy utilization efficiency, unstable power quality and other problems. Existing technologies have limitations in dealing with the dynamics and uncertainties of microgrids (such as the intermittent nature of renewable energy) and lack flexible and intelligent control strategies. Microgrid intelligent control methods are an evolving technical field involving distributed power generation, energy management, economic operation and other aspects. At present, the field of microgrid intelligent control methods is developing in a more efficient, intelligent, stable and environmentally friendly direction to meet the requirements of energy transformation and the digital era.

[0003] The "Photovoltaic Storage Charging and Inspection Integrated Intelligent Microgrid System and Its Control Method" involves the field of electric vehicle charging technology. It obtains the current and voltage output by the photovoltaic system in real time, calculates the photovoltaic power, and establishes a mathematical model of the energy storage system. The control method can generate a control strategy based on these data to achieve peak load shaving, reduce load power demand, and smooth the load curve. When the distribution network is out of power, the system can autonomously switch to the island operation mode to improve power supply reliability.

[0004] "Adaptive hierarchical control method and system for microgrid group with system power balancing" aims to achieve dynamic balancing of power in each sub-microgrid system. By adding a compensation control module and a power output balancing module on the basis of traditional droop control, the reference voltage and frequency of droop control can be dynamically corrected, and the influence of line impedance on the accuracy of active power distribution can be compensated. This control method can ensure the stability and efficiency of the microgrid group under different operating conditions."

[0005] "Coordination control device and coordination control method for smart microgrid" relates to the technical field of coordinated control of microgrid electricity consumption. The coordination control device of the smart microgrid is used to control the power supply unit, power consumption unit and circuit conduction, so that the working state of the microgrid is coordinated and stable.

[0006] "An Intelligent Control Method for Peak-to-Valley Difference of Electricity Consumption in Industrial and Commercial Parks" discloses an intelligent control method for peak-to-valley difference of electricity consumption in industrial and commercial parks, including: (1) constructing control targets and constraints for peak-to-valley difference of electricity consumption in industrial and commercial parks; (2) constructing a simplified mathematical model for real-time control; (3) modeling the continuous decision-making process of controlling the controllable load in the park in a scenario to achieve the control target as a Markov decision process, defining the state space and action space, and designing an immediate reward function; (4) based on the day-ahead load forecast curve, superimposing the forecast error probability distribution model, sampling and generating intelligent agent training and test data sets; (5) using the proximal strategy optimization method to perform distributed training on the reinforcement learning agent; (6) using the trained reinforcement learning agent for online intelligent control of peak-to-valley difference of electricity consumption in the park. By using the present invention, closed-loop autonomous training, updating and online real-time control of the reinforcement learning agent can be realized, which can effectively reduce the peak-to-valley difference and volatility of electricity consumption in industrial and commercial parks.

[0007] "A microgrid controller based on artificial intelligence control method" involves a microgrid controller based on artificial intelligence control method, which uses a historical sample database to form an observation set of a deep deterministic policy gradient network for action exploration, and then trains based on the continuous interaction between the microgrid controller and the microgrid system environment, directly constructing a mapping relationship between the microgrid operating conditions and the coordinated control results. The microgrid controller directly uses the offline trained microgrid controller model to make fast online decisions, with very fast calculation and response speeds, and regularly updates the offline training sample library to achieve rolling optimization of the model and continuous improvement of the strategy, realizing real-time coordinated control of the microgrid controller over the source, network, load, and storage within the system, effectively improving the efficiency and robustness of the microgrid controller.

[0008] Existing microgrid intelligent control methods use limited data, making it difficult to put renewable energy power generation forecast information, current environmental information, and load information into a model for decision-making, making it difficult to achieve control accuracy. Summary of the invention

[0009] The purpose of the present invention is to address the deficiencies of the prior art and provide a microgrid intelligent control method and device based on multi-agent collaboration.

[0010] The purpose of the present invention is to achieve the following technical solution: a microgrid intelligent control method based on multi-agent collaboration,

[0011] (1) We comprehensively collected various microgrid-related manuals, conducted in-depth research on professional books on microgrids, extracted relevant information on microgrid architecture, energy management, and stability analysis, carefully analyzed various microgrid cases, and sorted out various microgrid solutions to build a microgrid data and knowledge base;

[0012] (2) Extract key information related to microgrid control from microgrid data and knowledge base and train and fine-tune the general large model to obtain the microgrid control large model;

[0013] (3) Collect historical data on power generation, energy storage, and load usage to learn and train the deep time series model, respectively, to obtain a power generation prediction model, an energy storage prediction model, and a load prediction model, and use the power generation prediction model as a power generation agent, the energy storage prediction model as an energy storage agent, and the load prediction model as a load agent; a multi-agent system is constructed by constructing the power generation agent, the energy storage agent, and the load agent;

[0014] (4) Based on the power generation intelligent agent, energy storage intelligent agent, load intelligent agent and microgrid control large model, a microgrid intelligent control integrated model is constructed; the output of the power generation intelligent agent is used as the input of the microgrid intelligent control integrated model to provide prediction information for the energy production end of the microgrid; and the output of the load intelligent agent is used as the input of the microgrid intelligent control integrated model. The microgrid intelligent control integrated model arranges the distribution and use of energy according to the load prediction; the microgrid control large model comprehensively controls the various components in the microgrid based on the input of the power generation intelligent agent, energy storage intelligent agent, load intelligent agent and its own control logic.

[0015] Furthermore, the deep timing model is a Transformer model.

[0016] Further, the power generation intelligent agent corresponds to the power generation equipment in the microgrid and monitors the operating status of the power generation equipment;

[0017] The energy storage intelligent agent corresponds to the energy storage device in the microgrid and manages the charging and discharging process of the energy storage device;

[0018] The load intelligent agent corresponds to the load in the microgrid and monitors the change of the load demand;

[0019] In the multi-agent system, each agent exchanges information via a communication network. The power generation agent sends power generation prediction information to the energy storage agent and the load agent, and the load agent sends load demand prediction information to the power generation agent and the energy storage agent.

[0020] The present invention also includes a microgrid intelligent control device based on multi-agent collaboration, including a memory and one or more processors, wherein the memory stores executable code, and when the one or more processors execute the executable code, it is used for the above-mentioned microgrid intelligent control method based on multi-agent collaboration.

[0021] The present invention also includes a computer-readable storage medium on which a program is stored. When the program is executed by a processor, the above-mentioned microgrid intelligent control method based on multi-agent collaboration is implemented.

[0022] The beneficial effect of the present invention is that the present invention, by constructing a more comprehensive decision-making system, is conducive to the microgrid to achieve more intelligent and precise strategic control, thereby improving the economic benefits of multi-energy complementary synergy. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a flow chart of a microgrid intelligent control method based on multi-agent collaboration;

[0024] Figure 2 It is a logical architecture diagram;

[0025] Figure 3 This is a structural diagram of a microgrid intelligent control device based on multi-agent collaboration. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical scheme and advantages of the present invention more clear, the present invention is further described in detail in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0027] 1. Architecture of multi-agent system

[0028] The present invention constructs a multi-agent system, including multiple different types of agents. For example, a power generation agent (the power generation agent may include a photovoltaic agent and a wind power agent, the photovoltaic agent corresponds to the photovoltaic power generation equipment in the microgrid, and the wind power agent corresponds to the wind power generation equipment in the microgrid), an energy storage agent (the energy storage agent may correspond to the battery energy storage system in the microgrid) and a load agent (different load agents may be constructed, corresponding to different types of loads in the microgrid).

[0029] Each agent has its own attributes and functions. The power generation agent is responsible for monitoring the operating status of the power generation equipment (such as power generation, equipment health, etc.), the energy storage agent manages the charging and discharging process of the energy storage equipment, and the load agent monitors changes in load demand.

[0030] 2. Collaborative control mechanism of multi-agent system

[0031] The common control goal of the multi-agent system is to optimize the operation of the microgrid, including improving energy efficiency, maintaining power quality (such as stable voltage and frequency), and meeting load demand.

[0032] In a multi-agent system, each agent exchanges information through a specific communication network. For example, the power generation agent sends power generation forecast information (future power generation predicted based on weather forecasts, equipment performance, etc.) to the energy storage agent and the load agent.

[0033] The multi-agent coordination strategy includes the energy storage agent formulating charging and discharging strategies based on the load demand forecast (provided by the load agent) and the power generation forecast (provided by the power generation agent). For example, when there is excess power generation and low load demand, the energy storage agent starts the charging mode; when there is insufficient power generation and high load demand, the energy storage agent discharges to supplement the power gap.

[0034] Power generation agents also collaborate with each other. For example, power allocation is adjusted between different power generation equipment (photovoltaic and wind power) based on resource availability and economy. If the cost of photovoltaic power generation is low and resources are sufficient, and certain conditions are met, the photovoltaic power generation power is appropriately increased and the wind power power is reduced (taking into account factors such as the intermittent nature of wind power).

[0035] 3. Intelligent decision-making algorithm based on collaboration between large and small models

[0036] In the present invention, each intelligent agent is used as a small model, and the microgrid control model is used as a large model. The real-time status of each intelligent agent is obtained based on the microgrid control model to realize the decision-making process. The microgrid control model calculates the optimal control strategy through the microgrid control model based on the environmental information obtained from the environment, and the environmental information includes the current power generation status, load status and energy storage status. With the microgrid control model as the base, the execution process of the power generation, load and energy storage intelligent agents is controlled through the large model base, and the model parameters can be intermittently updated according to the actual environmental data, so that the control strategy is gradually optimized in the actual operation of the microgrid.

[0037] Example 1

[0038] like Figure 1 As shown, the present invention provides a microgrid intelligent control method based on multi-agent collaboration, comprising the following steps:

[0039] (1) Comprehensively collect various microgrid-related manuals; these manuals cover all aspects of microgrids from basic theory to practical operation, such as detailed parameters of different types of microgrid components, principles of various control strategies, etc.

[0040] Study professional books on microgrids in depth and extract important knowledge about microgrid architecture, energy management, stability analysis, etc.

[0041] Carefully analyze a large number of microgrid cases. Whether they are successful examples or problematic cases, they can provide valuable experience. For example, the innovative component combination methods and efficient energy scheduling solutions in some successful cases, while the problems and improvement directions in failed cases are also worth learning from.

[0042] We sorted out various microgrid solutions, including design solutions, construction solutions, and operation solutions for different application scenarios (such as power supply in remote areas, microgrids in industrial parks, etc.). By integrating the information in these manuals, books, cases, and solutions, we built a comprehensive and detailed microgrid data and knowledge base to provide data support and theoretical basis for subsequent work.

[0043] (2) Extract key information related to microgrid control from microgrid data and knowledge base, such as the optimal operating parameters of different components and control logic under different scenarios.

[0044] This information can be used to fine-tune the microgrid control model. For example, if the knowledge base shows that a certain type of energy storage device has the best charging and discharging efficiency within a specific power range, the corresponding control strategy can be adjusted in the large model to make the control of the energy storage device more precise.

[0045] For example, if the knowledge base shows that a certain type of energy storage device has the best charging and discharging efficiency within a specific power range, then the corresponding control strategy is adjusted in the large model to make the control of the energy storage device more precise.

[0046] (2) Extract key information related to microgrid control from the microgrid data and knowledge base, and train and fine-tune the general large model to obtain the microgrid control large model.

[0047] In terms of data collection, for wind power generation, it is necessary to obtain long-term historical data in different geographical locations and different types of wind turbines, including wind speed, wind direction, wind power generation, etc. The time span of these data should be as long as possible to cover the impact of various climate conditions and seasonal changes.

[0048] For photovoltaic power generation, historical data such as light intensity, light duration, photovoltaic panel temperature, and photovoltaic power generation are collected. These data sources can be monitoring systems of multiple photovoltaic power stations to ensure data accuracy and diversity.

[0049] The collection of load usage data involves data such as power consumption and time of different types of users (residential, industrial, commercial, etc.) in different time periods (peak, valley, flat, etc.). Data is collected through various means such as smart meters and power monitoring systems.

[0050] Based on the collected data, a deep time series model such as the Transformer model is selected. The Transformer model is selected because of its advantages in processing long sequence data. For the wind power prediction model, the data related to wind power generation is input into the model. Through the learning and training of the model, it can accurately predict the future wind power generation, taking into account the fluctuation of wind speed, the performance changes of wind turbines and other factors.

[0051] When building a photovoltaic prediction model, the historical data of photovoltaic power generation is used to allow the model to learn the complex relationship between sunlight and power generation, while considering the impact of factors such as temperature on power generation efficiency, so that photovoltaic power generation can be predicted under different weather and seasons.

[0052] The load forecasting model analyzes the historical data of load usage, taking into account factors such as user electricity usage habits, holidays, and seasonal changes, and accurately predicts the electricity load in various time periods in the future. These constructed forecasting models, as small model intelligent entities, each undertakes the forecasting tasks of wind power, photovoltaic power, and load, and provides basic forecasting information for the intelligent control of microgrids.

[0053] (3) Collect historical data on power generation, energy storage, and load usage to learn and train the deep time series model respectively, and obtain the power generation prediction model, energy storage prediction model, and load prediction model. The power generation prediction model is used as the power generation agent, the energy storage prediction model is used as the energy storage agent, and the load prediction model is used as the load agent. A multi-agent system is constructed by constructing the power generation agent, the energy storage agent, and the load agent.

[0054] (4) Based on the power generation intelligent agent, energy storage intelligent agent, load intelligent agent and microgrid control large model, a microgrid intelligent control integrated model is constructed; the output of the power generation intelligent agent is used as the input of the microgrid intelligent control integrated model to provide prediction information for the energy production end of the microgrid so as to make reasonable scheduling and control decisions in advance; and the output of the load intelligent agent is used as the input of the microgrid intelligent control integrated model, and the microgrid intelligent control integrated model arranges the distribution and use of energy according to the load prediction; the microgrid control large model comprehensively controls the various components in the microgrid (such as wind turbines, photovoltaic panels, energy storage equipment, load switches, etc.) according to the input of the power generation intelligent agent, energy storage intelligent agent, load intelligent agent and its own control logic.

[0055] At the same time, the compatibility between models and the smoothness of data interaction should be considered during the construction process to ensure that the entire microgrid intelligent control integrated model can operate stably and efficiently and achieve the intelligent control goals of the microgrid, such as improving energy utilization efficiency and enhancing power supply reliability.

[0056] According to a microgrid intelligent control method based on multi-agent collaboration provided in Example 1, its logical architecture is as follows Figure 2 shown.

[0057] like Figure 2 As shown, it can be divided into 4 layers from bottom to top:

[0058] The first layer is the management of data and knowledge such as materials, manuals, cases, etc. related to microgrids, as well as the management of general large model bases, such as the Zhipu large model.

[0059] The second layer is mainly a large microgrid control model. As the "brain" of microgrid control, it can regulate multiple small intelligent agent models in the upper layer.

[0060] The third layer is a variety of small model intelligent agent models such as power generation, load, diagnosis, etc., to achieve intelligent prediction and collaboration of corresponding businesses.

[0061] The fourth layer is the human-computer interaction layer, which mainly realizes question and answer, intelligent statistical analysis of data, etc. through the form of intelligent assistants.

[0062] Based on the above logical architecture diagram, the specific implementation steps are as follows:

[0063] (a.1) System construction and initialization

[0064] Component identification and modeling: Detailed parameter identification and mathematical modeling of various components in the microgrid, such as distributed power sources (solar panels, wind turbines, etc.), energy storage devices (batteries, supercapacitors, etc.), and loads (industrial loads, residential loads, etc.). Determine the basic architecture and functional scope of large and small models. The large model can be responsible for macroscopic system status assessment and long-term strategy planning, while the small model focuses on precise control of local components and short-term status prediction.

[0065] Multi-agent setting: Assign agents to different components or functional modules in the microgrid. For example, assign an agent to each distributed power source, an agent to each energy storage device, and loads can also be assigned agents based on type or area. Define the basic properties of each agent, including the range of devices it can control, communication protocol, initial state, etc. For example, the agent responsible for solar panels can control the corresponding panels, the communication protocol uses a specific power line communication protocol, and the initial state is the actual output power and operating temperature of the current panels.

[0066] (a.2) Data collection and preprocessing

[0067] Sensor deployment and data collection: Deploy sensors at key locations of the microgrid, such as at the output end of the distributed power supply, the charging and discharging port of the energy storage device, and the load access point. Sensor types include current sensors, voltage sensors, temperature sensors, etc. Collect various types of data, such as voltage, current, power, temperature, light intensity (for solar panels), wind speed (for wind turbines), etc. Collect data at a certain time interval (such as once every 10 seconds).

[0068] Data preprocessing: Clean the collected data and remove outliers. For example, if the collected voltage value exceeds the normal range (such as exceeding 120% of the rated voltage or less than 80%), it is judged as an outlier and corrected or deleted. Normalize the data and convert data of different magnitudes to the same magnitude range to facilitate subsequent model processing.

[0069] (a.3) Collaborative work between large and small models

[0070] Small model local analysis and prediction: Each small model performs local analysis based on the data of the component or area it is responsible for. For example, the small model responsible for the solar panel predicts the power output change trend of the panel in the short term (such as 1-5 minutes) in the future based on the collected light intensity and temperature data combined with the mathematical model of the panel. The small model passes the local analysis results and prediction information to the corresponding agent. For example, if it is predicted that the power of the solar panel will drop by 10% in the next 2 minutes, this information will be passed to the agent responsible for the panel.

[0071] Global evaluation and strategy formulation of large models: The large model collects local information transmitted by each small model and global data of the entire microgrid, such as total load demand, total capacity and current status of energy storage equipment. Perform a global system status assessment, such as judging the power supply reliability and energy balance status of the current microgrid. If it is found that the total load demand is about to exceed the supply capacity of the current distributed power supply and energy storage equipment, it is determined that the power supply reliability is reduced. Formulate macro strategies based on the global evaluation results. For example, formulate a charging and discharging strategy for energy storage equipment, such as formulating a charging strategy for energy storage equipment during low load periods (such as at night) and when the distributed power supply has surplus power; formulate a discharging strategy for energy storage equipment during peak load periods and when the distributed power supply is insufficient.

[0072] (a.4) Multi-agent collaborative execution

[0073] Agent communication and collaboration: Agents communicate and collaborate with each other according to the strategies formulated by the large model. For example, the agent responsible for energy storage equipment communicates with the agent responsible for distributed power supply to inform it of the current charging and discharging status and capacity of the energy storage equipment so that the distributed power supply agent can adjust the output power. Agents negotiate to determine their specific operational tasks. For example, during the load adjustment process, the load agent negotiates with the distributed power supply agent to determine whether to meet the energy balance needs of the system by reducing the load power (such as controlling the interruptible load) or increasing the output power of the distributed power supply (such as increasing the speed of the wind turbine to increase the power output).

[0074] Execution and feedback: The agent performs operations according to the tasks determined by negotiation. For example, the energy storage device agent performs charging or discharging operations according to the strategy formulated by the large model, which is achieved by controlling the switches and power regulators of the charging circuit or the discharging circuit. After the operation is executed, the agent feeds back the execution results to the large model and the small model. For example, the energy storage device agent feeds back the actual charging and discharging power, power change and other data to the large model and the small model responsible for the device for the next round of control adjustments.

[0075] (a.5) System monitoring and optimization

[0076] Real-time monitoring: Continuously monitor the operating status of the microgrid, including the operating parameters of each component, the execution status of the intelligent body, etc. For example, real-time monitoring of the battery health status of the energy storage device (by monitoring the internal resistance, voltage change and other parameters of the battery), the output stability of the distributed power supply, etc.

[0077] Optimization and adjustment: Based on the monitoring results, the large model and the small model optimize and adjust the control strategy. For example, if the battery health status of the energy storage device is found to be declining rapidly, the large model adjusts the charging and discharging strategy of the energy storage device, such as reducing the charging and discharging power or adjusting the charging and discharging time interval to extend the battery life; the small model adjusts its local control parameters according to the actual operation of the component, such as adjusting certain parameters in the maximum power tracking algorithm of the wind turbine to improve the power generation efficiency.

[0078] Example 2

[0079] The present embodiment relates to a microgrid intelligent control device based on multi-agent collaboration, including a memory and one or more processors, wherein the memory stores executable code, and when the one or more processors execute the executable code, it is used for a microgrid intelligent control method based on multi-agent collaboration in the above-mentioned embodiment 1; the device embodiment can be applied to any device with data processing capabilities, and the any device with data processing capabilities can be a device or apparatus such as a computer.

[0080] like Figure 3At the hardware level, the model watermark device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 1 Of course, in addition to the software implementation, the present invention does not exclude other implementations, such as logic devices or a combination of software and hardware, etc., that is, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0081] For the improvement of a technology, it can be clearly distinguished whether it is a hardware improvement (for example, improvement of the circuit structure of diodes, transistors, switches, etc.) or a software improvement (improvement of the method flow). However, with the development of technology, many improvements of the method flow today can be regarded as direct improvements of the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that the improvement of a method flow cannot be implemented with a hardware entity module. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming themselves, without having to ask chip manufacturers to design and make dedicated integrated circuit chips. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages ​​and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

[0082] The controller can be implemented in any appropriate manner, for example, the controller can take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (such as software or firmware) that can be executed by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art also know that in addition to implementing the controller in a purely computer-readable program code manner, the controller can be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, this controller can be considered as a hardware component, and the devices included therein for implementing various functions can also be regarded as structures within the hardware component. Or even, the devices for implementing various functions can be regarded as both software modules for implementing the method and structures within the hardware component.

[0083] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0084] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0085] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0086] The present invention may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0087] Example 3

[0088] An embodiment of the present invention further provides a computer-readable storage medium having a program stored thereon. When the program is executed by a processor, a microgrid intelligent control method based on multi-agent collaboration of the above-mentioned embodiment 1 is implemented.

[0089] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A microgrid intelligent control method based on multi-agent collaboration, characterized in that: The following steps are involved: (1) We comprehensively collected various microgrid-related manuals, conducted in-depth research on professional books on microgrids, extracted relevant information on microgrid architecture, energy management, and stability analysis, carefully analyzed various microgrid cases, and sorted out various microgrid solutions to build a microgrid data and knowledge base; (2) Extract key information related to microgrid control from microgrid data and knowledge base and train and fine-tune the general large model to obtain the microgrid control large model; (3) Collect historical data on power generation, energy storage, and load usage to learn and train the deep time series model, respectively, to obtain a power generation prediction model, an energy storage prediction model, and a load prediction model, and use the power generation prediction model as a power generation agent, the energy storage prediction model as an energy storage agent, and the load prediction model as a load agent; a multi-agent system is constructed by constructing the power generation agent, the energy storage agent, and the load agent; (4) Based on the power generation intelligent agent, energy storage intelligent agent, load intelligent agent and microgrid control large model, a microgrid intelligent control integrated model is constructed; the output of the power generation intelligent agent is used as the input of the microgrid intelligent control integrated model to provide prediction information for the energy production end of the microgrid; and the output of the load intelligent agent is used as the input of the microgrid intelligent control integrated model. The microgrid intelligent control integrated model arranges the distribution and use of energy according to the load prediction; the microgrid control large model comprehensively controls the various components in the microgrid based on the input of the power generation intelligent agent, energy storage intelligent agent, load intelligent agent and its own control logic.

2. According to claim 1, a microgrid intelligent control method based on multi-agent collaboration is characterized in that: The deep time series model is a Transformer model.

3. According to claim 1, a microgrid intelligent control method based on multi-agent collaboration is characterized in that: The power generation intelligent agent corresponds to the power generation equipment in the microgrid and monitors the operating status of the power generation equipment; The energy storage intelligent agent corresponds to the energy storage device in the microgrid and manages the charging and discharging process of the energy storage device; The load intelligent agent corresponds to the load in the microgrid and monitors the change of the load demand; In the multi-agent system, each agent exchanges information through a communication network. The power generation agent sends power generation prediction information to the energy storage agent and the load agent, and the load agent sends load demand prediction information to the power generation agent and the energy storage agent.

4. A microgrid intelligent control device based on multi-agent collaboration, characterized in that: It comprises a memory and one or more processors, wherein the memory stores executable code, and when the one or more processors execute the executable code, it is used to implement a microgrid intelligent control method based on multi-agent collaboration as described in any one of claims 1 to 3.

5. A computer-readable storage medium, characterized in that: A program is stored thereon, and when the program is executed by a processor, a microgrid intelligent control method based on multi-agent collaboration as described in any one of claims 1 to 3 is implemented.

Images