Multi-agent-based virtual power plant dispatching instruction rapid decomposition method

Abstract

The invention belongs to the technical field of virtual power plant instruction decomposition, and particularly relates to a virtual power plant dispatching instruction rapid decomposition method based on multiple agents. According to the method, the change and demand fluctuation of the power system can be more flexibly coped with, the compensation value is calculated based on the historical data and the current state of the intelligent agent, reasonable distribution and optimal use of system resources, loads and power are ensured, the overall operation efficiency of the virtual power plant is improved, and through the setting of the instruction decomposition level, the operation efficiency is improved. Task allocation of the intelligent agents can be dynamically adjusted according to actual conditions, overload of a single intelligent agent is avoided, execution efficiency of scheduling instructions is improved, high-energy-consumption tasks can be preferentially arranged when power is low by combining current power data, or generating capacity and operation cost are reduced when the power is high, economic benefits are improved, and the method is suitable for large-scale popularization and application. Compensation value calculation and decision making are carried out by utilizing historical data of the intelligent agent, so that the scientificity and accuracy of dispatching instruction decomposition are improved, and the influence of uncertain factors in the dispatching process is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of virtual power plant instruction decomposition, and in particular relates to a method for quickly decomposing virtual power plant dispatching instructions based on multi-agents. Background Art

[0002] With the transformation of the global energy structure and the rapid development of renewable energy, virtual power plants integrate decentralized distributed energy, energy storage systems, controllable loads and other resources through advanced information communication technology and control technology to form a unified, coordinated and controllable virtual power source. Virtual power plants can achieve optimal scheduling, supply and demand balance and collaborative management of energy, improve energy utilization efficiency and the stability of the power system. In the scheduling process of virtual power plants, the decomposition of scheduling instructions is one of the key links. The decomposition speed of scheduling instructions directly affects the response speed and scheduling effect of virtual power plants.

[0003] Traditional scheduling methods usually rely on centralized optimization calculations. With the expansion of the scale of virtual power plants and the increase in the number of participating distributed energy sources, the complexity and calculation time of centralized calculations have increased significantly, making it difficult to meet the requirements of real-time and high efficiency. In the process of scheduling instruction allocation, unbalanced resource utilization is prone to occur, with some intelligent agents being overloaded while other intelligent agents are idle, resulting in reduced overall system operating efficiency. Summary of the invention

[0004] The purpose of the present invention is to provide a method for quickly decomposing virtual power plant dispatch instructions based on multi-agents, which can effectively improve the operating efficiency and system stability of the virtual power plant by optimizing the decomposition and execution process of dispatch instructions.

[0005] The technical solution adopted by the present invention is as follows:

[0006] A multi-agent-based virtual power plant dispatch instruction fast decomposition method, comprising:

[0007] Obtaining dispatch instruction data in the virtual power plant, and obtaining multiple sub-instructions according to the dispatch instruction data;

[0008] Get the agent that executes each sub-command;

[0009] Get the current power data of each agent;

[0010] Obtain the historical system resource usage data of each agent when executing instructions, and obtain the system resource compensation value based on the historical system resource usage data;

[0011] Obtain the historical load data of each intelligent agent, and obtain the load compensation value according to the historical load data;

[0012] Obtaining power demand data and power generation capacity data of each intelligent entity, and obtaining a power compensation value according to the power demand data and power generation capacity data;

[0013] Obtaining the instruction decomposition level according to the current power data, the system resource compensation value, the load compensation value and the power compensation value;

[0014] The scheduling instruction is decomposed into multiple levels of sub-instructions according to the instruction decomposition level, and multiple agents are made to execute the multiple levels of sub-instructions corresponding to the decomposition in sequence according to the instruction decomposition level.

[0015] In a preferred solution, the step of obtaining the historical system resource usage data of each agent when executing instructions, and obtaining the system resource compensation value according to the historical system resource usage data, includes:

[0016] Get historical system resource usage data for each agent when executing instructions;

[0017] According to the historical system resource usage data, a plurality of corresponding historical system resource usage rates, serial numbers of the historical system resource usage rates and a total number of the historical system resource usage rates are obtained;

[0018] Get standard system resource usage;

[0019] The system resource compensation value is calculated according to multiple historical system resource usages, the numbers of the historical system resource usages, the total number of the historical system resource usages and the standard system resource usage, wherein the calculation formula is:

[0020]

[0021] In the formula, C B It is represented by the system resource compensation value, n is the total number of historical system resource usage, i is the number of historical system resource usage, C i Expressed as the i-th historical system resource usage, C b Expressed as standard system resource usage.

[0022] In a preferred solution, the step of obtaining the historical load data of each intelligent agent and obtaining the load compensation value according to the historical load data includes:

[0023] Get the historical load data of each agent;

[0024] According to the historical load data, a plurality of corresponding historical load values, serial numbers of the historical load values ​​and the total number of the historical load values ​​are obtained;

[0025] The load compensation value is calculated according to a plurality of historical load values, the numbers of the historical load values ​​and the total number of the historical load values, wherein the calculation formula is:

[0026]

[0027] Where P B represents the load compensation value, x represents the number of the historical load value, y represents the total number of historical load values, P x Expressed as the xth historical load value, P x-1 Expressed as the x-1th historical load value.

[0028] In a preferred embodiment, the step of obtaining the power demand data and power generation capacity data of each intelligent entity and obtaining the power compensation value according to the power demand data and power generation capacity data includes:

[0029] Obtain the power demand data and power generation capacity data of each intelligent agent;

[0030] Acquire multiple corresponding power demand values, the total number of power demand values ​​and the numbers of the power demand values ​​according to the power demand data;

[0031] Acquire multiple corresponding power generation capacity values, the total number of power generation capacity values ​​and the numbers of the power generation capacity values ​​according to the power generation capacity data;

[0032] The power compensation value is calculated according to a plurality of power demand values, the total number of power demand values, the numbers of the power demand values, a plurality of power generation capacity values, the total number of power generation capacity values, and the numbers of the power generation capacity values, wherein the calculation formula is:

[0033]

[0034] Where D B It is represented by the power compensation value, f is represented by the sum of the power generation capacity value and the power demand value, E g Expressed as the g-th power demand value, A g Expressed as the g-th power generation capacity value.

[0035] In a preferred embodiment, the step of obtaining the instruction decomposition level according to the current power data, the system resource compensation value, the load compensation value and the power compensation value comprises:

[0036] Calculate the instruction decomposition value according to the current power data, the system resource compensation value, the load compensation value and the power compensation value;

[0037] Get the instruction decomposition level according to the instruction decomposition value.

[0038] In a preferred embodiment, the step of calculating the instruction decomposition value according to the current power data, the system resource compensation value, the load compensation value and the power compensation value comprises:

[0039] Obtain the corresponding current power value according to the current power data;

[0040] The instruction decomposition value is calculated according to the current power value, the system resource compensation value, the load compensation value and the power compensation value, wherein the calculation formula is:

[0041] Z=Y·C B ·P B ·D B ;

[0042] In the formula, Z represents the command decomposition value, Y represents the current power value, C B Expressed as system resource compensation value, P B Expressed as load compensation value, D B Expressed as power compensation value.

[0043] In a preferred solution, the step of obtaining the instruction decomposition level according to the instruction decomposition value includes:

[0044] Obtaining an instruction level table, wherein the instruction level table includes a plurality of instruction decomposition interval values ​​and an instruction decomposition level corresponding to each instruction decomposition interval value;

[0045] Obtain a target instruction decomposition interval value according to the instruction decomposition value;

[0046] The instruction decomposition level is obtained from the instruction level table according to the target instruction decomposition interval value.

[0047] In a preferred embodiment, after the step of obtaining the instruction decomposition level from the instruction level table according to the target instruction decomposition interval value, the method further comprises:

[0048] Get the threshold number of sub-instructions within the instruction decomposition level;

[0049] Determine whether the number of sub-instructions in each instruction decomposition level exceeds a threshold number of sub-instructions;

[0050] If the number of sub-instructions within the instruction decomposition level exceeds the sub-instruction threshold number, it is determined that the instruction decomposition level for executing the sub-instructions exceeds the normal execution number, marked as exceeding the instruction decomposition level, and a warning message is issued;

[0051] If the number of sub-instructions within the instruction decomposition level does not exceed the sub-instruction threshold number, it is determined that the instruction decomposition level for executing the sub-instructions is normal.

[0052] In a preferred embodiment, if the number of sub-instructions in the instruction decomposition level exceeds the sub-instruction threshold number, the step of determining that the instruction decomposition level of the sub-instruction to be executed exceeds the normal execution number, marking it as exceeding the instruction decomposition level, and issuing a warning message, further includes:

[0053] Build level adjustment cycle;

[0054] Obtain the number of times that the same instruction decomposition level is marked as exceeding the instruction decomposition level within the level adjustment cycle, and mark it as the same number of exceeding times;

[0055] Get the same number of times exceeding the threshold;

[0056] Determine whether the same exceeding times exceeds the same exceeding threshold times;

[0057] If the same number of exceeding times exceeds the same number of exceeding thresholds, it is determined that the instruction decomposition interval value corresponding to the instruction decomposition level is abnormally divided, and the instruction decomposition interval value corresponding to the instruction decomposition level is re-divided;

[0058] If the same number of exceeding times does not exceed the same number of exceeding thresholds, it is determined that the instruction decomposition interval value division corresponding to this instruction decomposition level is normal.

[0059] And, a multi-agent-based virtual power plant dispatch instruction fast decomposition terminal, comprising:

[0060] one or more processors;

[0061] a storage device having one or more programs stored thereon;

[0062] When the one or more programs are executed by the one or more processors, the one or more processors implement the multi-agent-based virtual power plant scheduling instruction rapid decomposition method.

[0063] The technical effects achieved by the present invention are:

[0064] The present invention realizes the rapid decomposition and execution of dispatch instructions through the collaborative work of multiple agents, can respond to changes in the power system and demand fluctuations more flexibly, calculates compensation values ​​based on the historical data and current status of the agents, ensures the reasonable allocation and optimal use of system resources, loads and electricity, and improves the overall operating efficiency of the virtual power plant. By setting the instruction decomposition level, the task allocation of the agents can be dynamically adjusted according to actual conditions to avoid overloading of a single agent and improve the execution efficiency of the dispatch instructions. Combined with the current power data, high-energy-consuming tasks can be given priority when the power is low, or the power generation can be reduced when the power is high, thereby reducing operating costs and improving economic benefits. The historical data of the agents is used to calculate and make decisions on compensation values, which improves the scientificity and accuracy of the decomposition of dispatch instructions and reduces the impact of uncertain factors in the dispatch process. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 It is a flow chart of the method provided by the present invention. DETAILED DESCRIPTION

[0066] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0067] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and 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.

[0068] Secondly, the term "one embodiment" or "embodiment" as used 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 a preferred embodiment" that appears in different places in this specification does not refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.

[0069] Secondly, the present invention is described in detail in conjunction with schematic diagrams. When describing the embodiments of the present invention in detail, for the convenience of explanation, the schematic diagrams are only examples and should not limit the scope of protection of the present invention.

[0070] Please see attached Figure 1 As shown, a method for fast decomposition of virtual power plant dispatch instructions based on multi-agent is provided, including:

[0071] S1. Obtaining dispatch instruction data in the virtual power plant, and obtaining multiple sub-instructions according to the dispatch instruction data;

[0072] S2, obtain the agent that executes each sub-command;

[0073] S3, obtaining the current power data of each agent;

[0074] S4, obtaining the historical system resource usage data of each agent when executing the instruction, and obtaining the system resource compensation value according to the historical system resource usage data;

[0075] S5. Obtain historical load data of each intelligent agent, and obtain a load compensation value according to the historical load data;

[0076] S6, obtaining power demand data and power generation capacity data of each intelligent body, and obtaining a power compensation value according to the power demand data and power generation capacity data;

[0077] S7, obtaining an instruction decomposition level according to current power data, system resource compensation value, load compensation value and power compensation value;

[0078] S8. Decompose the scheduling instruction into multiple levels of sub-instructions according to the instruction decomposition level, and enable multiple agents to execute the multiple levels of sub-instructions corresponding to the decomposition in sequence according to the instruction decomposition level.

[0079] As in the above steps S1 to S8, the dispatch instruction data in the virtual power plant usually includes an overall power dispatch plan. By analyzing these dispatch instructions, they are decomposed into multiple specific sub-instructions. Each sub-instruction corresponds to a different task, such as power generation, load adjustment, etc. Each intelligent agent in the virtual power plant can be a power device, a power generation unit, or one or more combination units of a load unit. The intelligent agent that executes each sub-instruction is obtained, and their current power data, historical system resource usage data, historical load data, power demand data, and power generation capacity data are collected. The system resource compensation value is calculated based on the historical system resource usage data (such as CPU, memory, bandwidth, etc.) of the intelligent agent when executing the instruction, and its ability and consumption to execute the instruction are evaluated. The load compensation value is calculated based on the historical load data of the intelligent agent (such as the power load curve, etc.), and its load response capability is evaluated. The power compensation value is calculated based on the power demand data and power generation capacity data of the intelligent agent, and its power supply and demand balancing capability is evaluated. Combined with the current power data, system resource compensation value, load compensation value, and power compensation value , comprehensively evaluate the execution capability of each intelligent agent, determine the instruction decomposition level, decompose the dispatch instruction into multiple levels of sub-instructions according to the determined instruction decomposition level, and assign them to each intelligent agent for execution in turn according to priority and execution capability. Through the collaborative work of multiple intelligent agents, the rapid decomposition and execution of dispatch instructions are realized, which can respond to changes in the power system and demand fluctuations more flexibly. Based on the historical data and current status of the intelligent agent, the compensation value is calculated to ensure the reasonable allocation and optimal use of system resources, load and electricity, and improve the overall operation efficiency of the virtual power plant. Through the setting of the instruction decomposition level, the task allocation of the intelligent agent can be dynamically adjusted according to the actual situation to avoid overloading of a single intelligent agent and improve the execution efficiency of the dispatch instruction. Combined with the current power data, high-energy-consuming tasks can be prioritized when the power is low, or the power generation can be reduced when the power is high, reducing operating costs and improving economic benefits. The historical data of the intelligent agent is used for compensation value calculation and decision-making, which improves the scientificity and accuracy of the dispatch instruction decomposition and reduces the impact of uncertain factors in the dispatch process.

[0080] In a specific embodiment, the step of obtaining the historical system resource usage data of each agent when executing instructions, and obtaining the system resource compensation value according to the historical system resource usage data, includes:

[0081] S401, obtaining historical system resource usage data when each agent executes instructions;

[0082] S402, acquiring a plurality of corresponding historical system resource usage rates, serial numbers of the historical system resource usage rates, and a total number of the historical system resource usage rates according to the historical system resource usage data;

[0083] S403, obtaining standard system resource usage rate;

[0084] S404, calculating the system resource compensation value according to the plurality of historical system resource usage rates, the serial numbers of the historical system resource usage rates, the total number of the historical system resource usage rates and the standard system resource usage rate, wherein the calculation formula is:

[0085]

[0086] In the formula, C B It is represented by the system resource compensation value, n is the total number of historical system resource usage, i is the number of historical system resource usage, C i Expressed as the i-th historical system resource usage, C b Expressed as standard system resource usage.

[0087] As in the above steps S401 to S404, each agent consumes certain system resources, such as CPU, memory, bandwidth, etc., when executing instructions. It is necessary to collect the resource usage data of these agents when executing similar instructions in the past for analysis and evaluation, and extract multiple historical system resource usage rates from the acquired historical data, for example, the CPU usage rate, memory usage rate, etc. of the agent in a specific time period, and record the numbers and totals of these historical system resource usage rates for subsequent calculation and analysis. The standard system resource usage rate is a pre-set benchmark value used to evaluate the resource utilization efficiency of the agent. This standard value can be based on industry standards, equipment performance indicators or system administrator settings. According to multiple historical system resource usage rates, their numbers and totals, and the standard system resource usage rate, the system resource compensation value is calculated. By analyzing the historical system usage rates of the agent in detail, the system resource compensation value is calculated. Resource usage data can accurately evaluate its resource consumption when performing different tasks, which helps to allocate scheduling instructions more accurately and avoid resource waste and overload. The calculation of system resource compensation value takes into account the difference between historical usage rate and standard usage rate, and can dynamically adjust the task allocation of intelligent agents. By understanding the resource usage history and current status of each intelligent agent, it can predict possible resource bottlenecks and take measures in advance to use historical data for calculations and decisions, reducing the influence of subjective factors in the scheduling process and improving the scientificity and objectivity of decision-making. Reasonable allocation and use of system resources can reduce operating costs and improve economic benefits. For example, by balancing the load of each intelligent agent, reducing resource consumption during peak hours, reducing energy costs, and adapting to environmental changes and demand fluctuations, this flexibility enables virtual power plants to better cope with uncertainty and improve response speed and adaptability.

[0088] In a specific embodiment, the step of obtaining the historical load data of each intelligent agent and obtaining the load compensation value according to the historical load data includes:

[0089] S501, obtaining historical load data of each intelligent agent;

[0090] S502, acquiring a plurality of corresponding historical load values, serial numbers of the historical load values, and the total number of the historical load values ​​according to the historical load data;

[0091] S503, calculating a load compensation value according to a plurality of historical load values, the serial numbers of the historical load values, and the total number of the historical load values, wherein the calculation formula is:

[0092]

[0093] Where P B represents the load compensation value, x represents the number of the historical load value, y represents the total number of historical load values, P x Expressed as the xth historical load value, P x-1 Expressed as the x-1th historical load value.

[0094] As in the above steps S501 to S503, each agent will have its load data recorded during its past operation. These data reflect the load conditions of the agent in different time periods, including power demand, output, etc. These historical load data need to be collected for further analysis and calculation. Multiple specific historical load values ​​are extracted from the acquired historical load data. These values ​​can represent the load conditions of the agent at different time points. The numbers and totals of these historical load values ​​are recorded for subsequent statistical analysis and calculation. The load compensation value is calculated based on the extracted multiple historical load values ​​and their numbers and totals. By analyzing the historical load data of the agent in detail, its response capability under different load conditions can be accurately evaluated, which helps to better Reasonable allocation of dispatching instructions ensures the stable operation of the power system. The calculation of load compensation value takes into account the historical changes of load data and can dynamically adjust the dispatching strategy according to the load characteristics of the intelligent agent. By analyzing historical load data, it can predict the load fluctuations that may occur when the intelligent agent executes instructions, and take measures to compensate in advance. The use of historical load data for calculation and decision-making reduces the influence of subjective factors in the dispatching process and improves the scientificity and objectivity of decision-making. Data-driven decisions can better reflect the actual situation, thereby optimizing the dispatching effect. By reasonably allocating and using load resources, operating costs can be reduced and economic benefits can be improved. For example, by balancing the loads of various intelligent agents, resource consumption during peak periods can be reduced, and energy costs can be reduced.

[0095] In a specific embodiment, the step of obtaining the power demand data and power generation capacity data of each intelligent entity and obtaining the power compensation value according to the power demand data and power generation capacity data includes:

[0096] S601, obtaining power demand data and power generation capacity data of each intelligent entity;

[0097] S602, acquiring a plurality of corresponding power demand values, a total number of power demand values, and numbers of the power demand values ​​according to the power demand data;

[0098] S603, acquiring a plurality of corresponding power generation capacity values, a total number of power generation capacity values, and numbers of the power generation capacity values ​​according to the power generation capacity data;

[0099] S604, calculating the power compensation value according to the plurality of power demand values, the total number of power demand values, the numbers of the power demand values, the plurality of power generation capacity values, the total number of power generation capacity values ​​and the numbers of the power generation capacity values, wherein the calculation formula is:

[0100]

[0101] Where D B It is represented by the power compensation value, f is represented by the sum of the power generation capacity value and the power demand value, E g Expressed as the g-th power demand value, A g Expressed as the g-th power generation capacity value.

[0102] As in the above steps S601 to S604, the power demand data and power generation capacity data of each intelligent entity are collected. These data reflect the power demand of the intelligent entity in different time periods and its power generation capacity. A plurality of specific power demand values ​​are extracted from the power demand data, and the numbers and total numbers of these demand values ​​are recorded. These values ​​represent the power demand of the intelligent entity at different time points. A plurality of specific power generation capacity values ​​are extracted from the power generation capacity data, and the numbers and total numbers of these power generation capacity values ​​are recorded. These values ​​represent the power generation capacity of the intelligent entity at different time points. The power compensation value is calculated based on the extracted power demand value and power generation capacity value. By analyzing the power demand and power generation capacity of each intelligent entity in detail, the power compensation value can be evaluated more accurately. Estimate its response capability under different load conditions. The calculation of the power compensation value takes into account the data of power demand and power generation capacity. It can dynamically adjust the power dispatch strategy according to the specific situation of the intelligent body, optimize resource allocation, reduce the influence of subjective factors, and improve the scientificity and objectivity of dispatch decisions. It can flexibly adjust the allocation of dispatch instructions to adapt to environmental changes and demand fluctuations. Reasonable allocation of power resources can reduce operating costs and improve economic benefits. For example, by balancing the power supply and demand of each intelligent body, reducing power waste and overload problems during peak periods, and reducing energy costs, through the analysis of power demand and power generation capacity data, it can predict future trends in power demand and power generation capacity changes, and help formulate more scientific dispatch plans.

[0103] In a specific embodiment, the step of obtaining the instruction decomposition level according to the current power data, the system resource compensation value, the load compensation value and the power compensation value includes:

[0104] S701, calculating the instruction decomposition value according to the current power data, the system resource compensation value, the load compensation value and the power compensation value;

[0105] S702. Obtain an instruction decomposition level according to the instruction decomposition value.

[0106] As in the above steps S701 to S702, current power data, system resource compensation value, load compensation value and power compensation value are collected. These data reflect the current market environment and the resource consumption and response capability of each agent when executing instructions. Based on these data, a comprehensive instruction decomposition value is calculated. Based on the calculated instruction decomposition value, the instruction decomposition level is determined. The instruction decomposition level can be divided into different levels according to the size of the decomposition value. Multiple thresholds can be set to divide the instruction decomposition value into different level intervals. Taking into account the current power and the resource usage of the agent, the scheduling instructions can be allocated more scientifically and reasonably to avoid the problem of incorrect instruction decomposition due to a single factor. The decision deviation caused by the decomposition value of the instruction can be calculated and the decomposition level can be determined according to it. The intelligent agents with high resource utilization efficiency and low execution cost can be given priority to execute the scheduling instructions, thereby optimizing the overall resource utilization efficiency. Combined with the current power data, the scheduling strategy can be dynamically adjusted to respond to market price changes in a timely manner, reduce operating costs, and improve economic benefits. By comprehensively evaluating the load compensation value and power compensation value of the intelligent agent, load changes and power demand can be better predicted and responded to. The setting of the instruction decomposition level makes it possible to flexibly adjust the execution order and priority of the scheduling instructions to adapt to different operating conditions and demand changes, thereby improving the flexibility and adaptability of scheduling.

[0107] In a specific embodiment, the step of calculating the instruction decomposition value according to the current power data, the system resource compensation value, the load compensation value and the power compensation value includes:

[0108] S7011. Obtain the corresponding current power value according to the current power data;

[0109] S7012. Calculate the instruction decomposition value according to the current power value, the system resource compensation value, the load compensation value and the power compensation value, wherein the calculation formula is:

[0110] Z=Y·C B ·P B ·D B ;

[0111] In the formula, Z represents the command decomposition value, Y represents the current power value, C B Expressed as system resource compensation value, P BExpressed as load compensation value, D B Expressed as power compensation value.

[0112] As in the above steps S7011 to S7012, the current power data is obtained from the market data or the internal power market system. The current power reflects the supply and demand situation of the power market and is an important reference data for making scheduling decisions. The corresponding current power value is calculated. The current power value can be real-time power, hourly average power or power in other time periods. According to the current power value, system resource compensation value, load compensation value and power compensation value, a comprehensive instruction decomposition value is calculated. This value will be used to determine the priority of the intelligent agent in executing the scheduling instruction. By comprehensively considering the current power data, system resource compensation value, load compensation value and power compensation value, the execution capability and cost of each intelligent agent can be evaluated more comprehensively and accurately to ensure the reasonable allocation of scheduling instructions. The introduction of current power data can respond to changes in the power market in a timely manner and give priority to scheduling high-energy consumption when power is low. tasks and reduce operating costs; reduce unnecessary energy consumption when power is high and improve economic benefits. By calculating the comprehensive instruction decomposition value, it can give priority to intelligent agents with high resource utilization efficiency and strong load response ability to execute scheduling instructions, optimize resource allocation, and improve overall operating efficiency. The comprehensive consideration of load compensation value and power compensation value can better predict and respond to load fluctuations and changes in power demand, and avoid system failures caused by load exceeding the ability of intelligent agents. Through data-based calculation and analysis, the subjective bias in scheduling decisions is reduced, the scientificity and objectivity of decisions are improved, and the scheduling strategy can be ensured to truly reflect the system status and market conditions. According to the real-time power data and the compensation value of each intelligent agent, the allocation of scheduling instructions can be flexibly adjusted to adapt to changes in demand under different time periods and market conditions, and the flexibility and adaptability of scheduling can be improved.

[0113] In a specific embodiment, the step of obtaining the instruction decomposition level according to the instruction decomposition value includes:

[0114] S7021, obtaining an instruction level table, wherein the instruction level table includes a plurality of instruction decomposition interval values ​​and an instruction decomposition level corresponding to each instruction decomposition interval value;

[0115] S7022, obtaining a target instruction decomposition interval value according to the instruction decomposition value;

[0116] S7023. Obtain the instruction decomposition level from the instruction level table according to the target instruction decomposition interval value.

[0117] As in the above steps S7021 to S7023, the instruction level table is a pre-set comparison table, which contains multiple instruction decomposition interval values ​​and their corresponding instruction decomposition levels. Each interval value represents a range, and each range corresponds to a specific level. The instruction decomposition level table can be set according to historical data and system optimization goals. According to the calculated instruction decomposition value, it is compared with the decomposition interval value in the instruction level table to determine which interval the value belongs to. For example, if the instruction decomposition value is 15, it falls within the range of 10-20. According to the target instruction decomposition interval value, the corresponding instruction decomposition level is found from the instruction level table. By setting a detailed instruction level table, Fine-grained management of different instruction decomposition values ​​ensures that the scheduling task allocation of each intelligent agent is more reasonable and accurate. The use of the instruction level table can simplify the operating process when facing complex data calculations and decisions, and facilitate the maintenance and update of scheduling strategies. Through the pre-set instruction decomposition interval values ​​and corresponding levels, the scheduling priority of each intelligent agent can be quickly determined to improve the efficiency and response speed of scheduling decisions. The instruction level table can be dynamically adjusted according to actual conditions to adapt to different market conditions and operating environments. The combination of instruction decomposition values ​​and instruction level tables ensures that scheduling decisions are based on scientific calculations and analysis results, reduces the influence of subjective factors, and improves the scientificity and rationality of decisions.

[0118] In a specific embodiment, after the step of obtaining the instruction decomposition level from the instruction level table according to the target instruction decomposition interval value, the method further includes:

[0119] S7024, obtaining the threshold number of sub-instructions within the instruction decomposition level;

[0120] S7025, determining whether the number of sub-instructions in each instruction decomposition level exceeds a sub-instruction threshold number;

[0121] If the number of sub-instructions within the instruction decomposition level exceeds the sub-instruction threshold number, it is determined that the instruction decomposition level for executing the sub-instructions exceeds the normal execution number, marked as exceeding the instruction decomposition level, and a warning message is issued;

[0122] If the number of sub-instructions within the instruction decomposition level does not exceed the sub-instruction threshold number, it is determined that the instruction decomposition level for executing the sub-instructions is normal.

[0123] As in the above steps S7024 to S7025, a sub-instruction threshold number is pre-set for each instruction decomposition level. This threshold represents the maximum number of sub-instructions that can be accommodated in each level under normal circumstances. This threshold can be set based on the historical data and operating experience of the system. For each instruction decomposition level, the number of sub-instructions currently allocated to the level is checked. If the number of sub-instructions in a certain level exceeds the preset threshold number, it is determined that the number of sub-instructions at the level exceeds the normal execution number, and it is marked as exceeding the instruction decomposition level, and a warning message is issued. If the number of sub-instructions in a certain level does not exceed the preset threshold number, it is determined that the number of sub-instructions at the level is within the normal range. By setting the sub-instruction threshold and performing real-time monitoring, It can effectively avoid overload operation caused by too many sub-instructions within a certain instruction decomposition level. When a certain instruction decomposition level exceeds the normal execution number, it can issue a warning message in time to remind the operation and maintenance personnel or system administrator to intervene. By monitoring the number of sub-instructions and setting thresholds, resources can be allocated and scheduled more reasonably to ensure the balanced operation of each intelligent body and avoid excessive use or waste of resources. Setting sub-instruction thresholds and real-time monitoring can help make scientific decisions based on actual operating conditions, ensure that the execution of scheduling instructions is within a controllable range, and improve the scientificity and rationality of scheduling strategies. By setting marks and warning messages that exceed the instruction decomposition level, an effective early warning mechanism is established, which can detect potential operating risks in advance and take measures to prevent them.

[0124] In a specific embodiment, if the number of sub-instructions in the instruction decomposition level exceeds the sub-instruction threshold number, then determining that the instruction decomposition level of the sub-instruction to be executed exceeds the normal execution number, marking it as exceeding the instruction decomposition level, and issuing a warning message, the step further includes:

[0125] S7026, build level adjustment cycle;

[0126] S7027, obtaining the number of times that the same instruction decomposition level is marked as exceeding the instruction decomposition level within the level adjustment period, and marking it as the same number of exceeding times;

[0127] S7028, obtaining the same number of times exceeding the threshold;

[0128] S7029, determining whether the same exceeding times exceeds the same exceeding threshold times;

[0129] If the same number of exceeding times exceeds the same number of exceeding thresholds, it is determined that the instruction decomposition interval value corresponding to the instruction decomposition level is abnormally divided, and the instruction decomposition interval value corresponding to the instruction decomposition level is re-divided;

[0130] If the same number of exceeding times does not exceed the same number of exceeding thresholds, it is determined that the instruction decomposition interval value division corresponding to this instruction decomposition level is normal.

[0131] As in the above steps S7026 to S7029, a fixed time period (for example, one day, half a day or one hour) is determined as the level adjustment period. During this period, the operation of the instruction decomposition level is continuously monitored and recorded. During each level adjustment period, the number of times the same instruction decomposition level is marked as exceeding the instruction decomposition level is counted. This number indicates the frequency of the number of sub-instructions in the level exceeding the preset threshold value within a period. A threshold is set to indicate the number of times the same instruction decomposition level is marked as exceeding the allowed number of instruction decomposition levels within a level adjustment period. If the number of exceeding times exceeds the same exceeding threshold number, it is determined that the interval value division of the instruction decomposition level is abnormal and is re-divided. If it does not exceed, it is determined that It is determined that the interval value division of the instruction decomposition level is normal. By monitoring the instruction decomposition level exceeding the threshold during the level adjustment cycle, the division of the instruction decomposition interval value can be dynamically adjusted and optimized to ensure its rationality and effectiveness. When a certain instruction decomposition level frequently exceeds the threshold, it can be timely identified and re-divided into interval values ​​to avoid long-term unreasonable interval division leading to unstable operation of the system. By setting the adjustment cycle and the number of times the threshold is exceeded, it can automatically monitor and adjust, reduce human intervention, and improve the efficiency and accuracy of scheduling management. At the same time, the establishment of an early warning mechanism can promptly remind system administrators to intervene and dynamically adjust the instruction decomposition interval value to ensure that the number of sub-instructions within each instruction decomposition level is evenly distributed to avoid excessive concentration or dispersion of resources.

[0132] And, a multi-agent-based virtual power plant dispatch instruction fast decomposition terminal, comprising:

[0133] one or more processors;

[0134] a storage device having one or more programs stored thereon;

[0135] When the one or more programs are executed by the one or more processors, the one or more processors implement the multi-agent-based virtual power plant scheduling instruction rapid decomposition method.

[0136] The above is only a preferred embodiment of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications should also be considered as the protection scope of the present invention. The structures, devices and operating methods not specifically described and explained in the present invention shall be implemented according to the conventional means in the art unless otherwise specified and limited.

Claims

1. A multi-agent-based virtual power plant dispatch instruction fast decomposition method, characterized in that: include: Obtaining dispatch instruction data in the virtual power plant, and obtaining multiple sub-instructions according to the dispatch instruction data; Get the agent that executes each sub-command; Get the current power data of each agent; Obtain the historical system resource usage data of each agent when executing instructions, and obtain the system resource compensation value based on the historical system resource usage data; Obtain the historical load data of each intelligent agent, and obtain the load compensation value according to the historical load data; Obtaining power demand data and power generation capacity data of each intelligent entity, and obtaining a power compensation value according to the power demand data and power generation capacity data; Obtaining the instruction decomposition level according to the current power data, the system resource compensation value, the load compensation value and the power compensation value; The scheduling instruction is decomposed into multiple levels of sub-instructions according to the instruction decomposition level, and multiple agents are made to execute the multiple levels of sub-instructions corresponding to the decomposition in sequence according to the instruction decomposition level.

2. The multi-agent-based virtual power plant dispatch instruction rapid decomposition method according to claim 1 is characterized in that: The step of obtaining the historical system resource usage data of each agent when executing instructions, and obtaining the system resource compensation value according to the historical system resource usage data, comprises: Get historical system resource usage data for each agent when executing instructions; According to the historical system resource usage data, a plurality of corresponding historical system resource usage rates, serial numbers of the historical system resource usage rates and a total number of the historical system resource usage rates are obtained; Get standard system resource usage; The system resource compensation value is calculated according to multiple historical system resource usages, the numbers of the historical system resource usages, the total number of the historical system resource usages and the standard system resource usage, wherein the calculation formula is: In the formula, C B It is represented by the system resource compensation value, n is the total number of historical system resource usage, i is the number of historical system resource usage, C i Expressed as the i-th historical system resource usage, C b Expressed as standard system resource usage.

3. The multi-agent-based virtual power plant dispatch instruction rapid decomposition method according to claim 1 is characterized in that: The step of obtaining the historical load data of each intelligent agent and obtaining the load compensation value according to the historical load data comprises: Get the historical load data of each agent; According to the historical load data, a plurality of corresponding historical load values, serial numbers of the historical load values ​​and the total number of the historical load values ​​are obtained; The load compensation value is calculated according to a plurality of historical load values, the numbers of the historical load values ​​and the total number of the historical load values, wherein the calculation formula is: Where P B represents the load compensation value, x represents the number of the historical load value, y represents the total number of historical load values, P x Expressed as the xth historical load value, P x-1 Expressed as the x-1th historical load value.

4. The multi-agent-based virtual power plant dispatch instruction rapid decomposition method according to claim 1 is characterized in that: The step of obtaining the power demand data and power generation capacity data of each intelligent entity and obtaining the power compensation value according to the power demand data and power generation capacity data includes: Obtain the power demand data and power generation capacity data of each intelligent agent; Acquire multiple corresponding power demand values, the total number of power demand values ​​and the numbers of the power demand values ​​according to the power demand data; Acquire multiple corresponding power generation capacity values, the total number of power generation capacity values ​​and the numbers of the power generation capacity values ​​according to the power generation capacity data; The power compensation value is calculated according to a plurality of power demand values, the total number of power demand values, the numbers of the power demand values, a plurality of power generation capacity values, the total number of power generation capacity values, and the numbers of the power generation capacity values, wherein the calculation formula is: Where D B It is represented by the power compensation value, f is represented by the sum of the power generation capacity value and the power demand value, E g Expressed as the g-th power demand value, A g Expressed as the g-th power generation capacity value.

5. The multi-agent-based virtual power plant dispatch instruction rapid decomposition method according to claim 1 is characterized in that: The step of obtaining the instruction decomposition level according to the current power data, the system resource compensation value, the load compensation value and the power compensation value comprises: Calculate the instruction decomposition value according to the current power data, the system resource compensation value, the load compensation value and the power compensation value; Get the instruction decomposition level according to the instruction decomposition value.

6. The multi-agent-based virtual power plant dispatch instruction rapid decomposition method according to claim 5 is characterized in that: The step of calculating the instruction decomposition value according to the current power data, the system resource compensation value, the load compensation value and the power compensation value comprises: Obtain the corresponding current power value according to the current power data; The instruction decomposition value is calculated according to the current power value, the system resource compensation value, the load compensation value and the power compensation value, wherein the calculation formula is: Z=Y·C B ·P B ·D B ; In the formula, Z represents the command decomposition value, Y represents the current power value, C B Expressed as system resource compensation value, P B Expressed as load compensation value, D B Expressed as power compensation value.

7. The multi-agent-based virtual power plant dispatch instruction rapid decomposition method according to claim 5 is characterized in that: The step of obtaining the instruction decomposition level according to the instruction decomposition value comprises: Obtaining an instruction level table, wherein the instruction level table includes a plurality of instruction decomposition interval values ​​and an instruction decomposition level corresponding to each instruction decomposition interval value; Obtain a target instruction decomposition interval value according to the instruction decomposition value; The instruction decomposition level is obtained from the instruction level table according to the target instruction decomposition interval value.

8. The multi-agent-based virtual power plant dispatch instruction rapid decomposition method according to claim 7 is characterized in that: After the step of obtaining the instruction decomposition level from the instruction level table according to the target instruction decomposition interval value, the method further includes: Get the threshold number of sub-instructions within the instruction decomposition level; Determine whether the number of sub-instructions in each instruction decomposition level exceeds a threshold number of sub-instructions; If the number of sub-instructions within the instruction decomposition level exceeds the sub-instruction threshold number, it is determined that the instruction decomposition level for executing the sub-instructions exceeds the normal execution number, marked as exceeding the instruction decomposition level, and a warning message is issued; If the number of sub-instructions within the instruction decomposition level does not exceed the sub-instruction threshold number, it is determined that the instruction decomposition level for executing the sub-instructions is normal.

9. The multi-agent-based virtual power plant dispatch instruction rapid decomposition method according to claim 8 is characterized in that: If the number of sub-instructions in the instruction decomposition level exceeds the sub-instruction threshold number, it is determined that the instruction decomposition level of the sub-instruction to be executed exceeds the normal execution number, marked as exceeding the instruction decomposition level, and issuing a warning message, the step also includes: Build level adjustment cycle; Obtain the number of times that the same instruction decomposition level is marked as exceeding the instruction decomposition level within the level adjustment cycle, and mark it as the same number of exceeding times; Get the same number of times exceeding the threshold; Determine whether the same exceeding times exceeds the same exceeding threshold times; If the same number of exceeding times exceeds the same number of exceeding thresholds, it is determined that the instruction decomposition interval value corresponding to the instruction decomposition level is abnormally divided, and the instruction decomposition interval value corresponding to the instruction decomposition level is re-divided; If the same number of exceeding times does not exceed the same number of exceeding thresholds, it is determined that the instruction decomposition interval value division corresponding to this instruction decomposition level is normal.

10. A multi-agent-based virtual power plant dispatch instruction rapid decomposition terminal, characterized in that: include: one or more processors; a storage device having one or more programs stored thereon; When the one or more programs are executed by the one or more processors, the one or more processors implement the multi-agent-based virtual power plant scheduling instruction rapid decomposition method as described in any one of claims 1-9.

Images