WO2016078329A1 - 一种多智能体结构的微电网优化运行方法 - Google Patents

一种多智能体结构的微电网优化运行方法 Download PDF

Info

Publication number
WO2016078329A1
WO2016078329A1 PCT/CN2015/077471 CN2015077471W WO2016078329A1 WO 2016078329 A1 WO2016078329 A1 WO 2016078329A1 CN 2015077471 W CN2015077471 W CN 2015077471W WO 2016078329 A1 WO2016078329 A1 WO 2016078329A1
Authority
WO
WIPO (PCT)
Prior art keywords
bidding
microgrid
agent
distributed power
unit
Prior art date
Application number
PCT/CN2015/077471
Other languages
English (en)
French (fr)
Inventor
孔祥玉
王晟晨
Original Assignee
天津大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 天津大学 filed Critical 天津大学
Priority to US15/526,738 priority Critical patent/US20190108600A1/en
Publication of WO2016078329A1 publication Critical patent/WO2016078329A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
    • G05B13/041—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a variable is automatically adjusted to optimise the performance
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks
    • H02J3/004—Generation forecast, e.g. methods or systems for forecasting future energy generation
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks
    • H02J3/008—Circuit arrangements for ac mains or ac distribution networks involving trading of energy or energy transmission rights
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks
    • H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
    • H02J3/381—Dispersed generators
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J2300/00—Systems for supplying or distributing electric power characterised by decentralized, dispersed, or local generation
    • H02J2300/10—The dispersed energy generation being of fossil origin, e.g. diesel generators
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J2300/00—Systems for supplying or distributing electric power characterised by decentralized, dispersed, or local generation
    • H02J2300/20—The dispersed energy generation being of renewable origin
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J2300/00—Systems for supplying or distributing electric power characterised by decentralized, dispersed, or local generation
    • H02J2300/20—The dispersed energy generation being of renewable origin
    • H02J2300/22—The renewable source being solar energy
    • H02J2300/24—The renewable source being solar energy of photovoltaic origin
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J2300/00—Systems for supplying or distributing electric power characterised by decentralized, dispersed, or local generation
    • H02J2300/20—The dispersed energy generation being of renewable origin
    • H02J2300/28—The renewable source being wind energy
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J2300/00—Systems for supplying or distributing electric power characterised by decentralized, dispersed, or local generation
    • H02J2300/30—The power source being a fuel cell
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J2310/00—The network for supplying or distributing electric power characterised by its spatial reach or by the load
    • H02J2310/10—The network having a local or delimited stationary reach
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E10/00—Energy generation through renewable energy sources
    • Y02E10/50—Photovoltaic [PV] energy
    • Y02E10/56—Power conversion systems, e.g. maximum power point trackers
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S50/00—Market activities related to the operation of systems integrating technologies related to power network operation or related to communication or information technologies
    • Y04S50/10—Energy trading, including energy flowing from end-user application to grid

Definitions

  • the invention relates to the field of power systems, in particular to a micro-grid optimization operation method of a multi-agent structure.
  • MAS Multi-Agent Systems
  • the micro-grid optimization operation using the MAS structure is aimed at the minimum cost based on the power generation, operation and maintenance costs of the network equipment or the cost of purchasing electricity from the main network [2], and the internal self-voluntary cooperation mode through the various agents (Agent) The full cooperation method for division of labor and cooperation.
  • This type of operation mode is simple and practical in the case of single property rights in the microgrid.
  • the microgrid marketization when there are different investment entities in the microgrid, such as the microgrid composed of many hotels, this type of operation cannot solve the problem of internal benefit distribution of the microgrid.
  • each DER in the microgrid is allowed to have its own goals and interests, and can participate in the operation of the microgrid in an intentional way. At this time, voluntary collaboration is only a special case of interaction between agents.
  • the bidding method and the microgrid control method can realize the decentralized decision of distributed power supply through the formulation of power generation bidding, and can also achieve centralized control by price coordination management. Due to the characteristics of the distributed power supply and the microgrid itself, the existing research results of the conventional power market cannot be simply transplanted into the microgrid, mainly in the following two aspects:
  • the distributed bidding function is added on the basis of the existing operation control of the microgrid, the interests of each microgrid participant are reflected by the market means, and the microgrid optimization operation is guided; the artificial immune system is utilized in the bidding process ( Artificial Immune systems (AIS)'s adaptive and defect tolerance capabilities deal with the uncertainty caused by intermittent power supplies; and through the co-evolution process of the technology, improve the coordination of the entire microgrid MAS.
  • Artificial Immune systems (AIS)'s adaptive and defect tolerance capabilities deal with the uncertainty caused by intermittent power supplies; and through the co-evolution process of the technology, improve the coordination of the entire microgrid MAS.
  • the invention provides a micro-grid optimization operation method with multi-agent structure, and the invention can solve the problem that the existing operation mode cannot realize the internal benefit distribution of the micro-grid when the equipment property rights in the micro-grid are not uniform or involve the public micro-grid. .
  • the method guides the optimization operation of the microgrid through the market mechanism of bidding. The specific process is described below:
  • a micro-grid optimization operation method of a multi-agent structure comprising the following steps:
  • microgrid distributed power source adopts unit bidding to realize quotation function; bid management agent assists in completing negotiation and optimization of each unit in microgrid;
  • the bidding management agent sends the bidding results of each unit to the microgrid control system, and the microgrid control system performs the optimal operation of the microgrid based on the results of the bidding negotiation.
  • the steps of the micro power grid distributed power source adopting the unit bidding to realize the quotation function are specifically:
  • the affinity formula of the artificial immune algorithm is determined to form an artificial immune quotation environment antigen
  • the antibody that meets the interests of the distributed power source is obtained, and the antibody is decoded to obtain the bidding scheme of the distributed power source, and submitted to the bid management agent.
  • the steps of the bid management agent to assist in completing the negotiation and optimization of each unit in the micro grid are as follows:
  • the objective function based on the distributed power supply itself bidding determines the affinity formula of the artificial immune algorithm, and the specific steps for forming the artificial immune quotation environment antigen are:
  • the distributed power source owner determines the quotation form and the target letter of its own bid based on the bidding model of the micro grid. number;
  • the bidding unit agent determines the antigenic representation of the artificial immune algorithm based on the characteristics of the distributed power source
  • the bidding unit agent determines the antibody of the artificial immune algorithm and the encoding and decoding formula of the antibody based on the quotation form of the distributed power source;
  • the bidding unit agent determines the affinity formula of the artificial immune algorithm based on the objective function of the distributed power source itself bidding;
  • the bidding unit agent processes the collected information to form an artificial immune quotation environment antigen.
  • the antigenic expression of the artificial immune algorithm includes the following three parts of information:
  • the judgment condition is: the specified number of times is reached, and/or the clearing price in the micro grid does not change any more.
  • the composition of the micro-grid may belong to different owners, and it is inevitable to optimize operation and control based on economic means such as bidding.
  • the invention can solve the problem that when the equipment property rights in the micro grid are not uniform, or the public micro grid is involved, the existing operation mode cannot distribute the internal benefits of the micro grid, and the realization of the commercial operation of the micro grid is facilitated.
  • the present invention adopts MAS's bid optimization operation and control, reflects the interests of each micro-grid participant through market means, and uses the bidding result to guide the micro-grid optimization, which is in line with the characteristics of micro-grid stratification and distributed control.
  • the programmable controller embedded in the algorithm according to the MAS standard in the manufacturing controller can realize “plug and play” and directly participate in the system bid optimization operation.
  • Figure 2 is a schematic diagram of the implementation time process of the microgrid optimization operation
  • FIG. 5 is a flow chart of an optimization algorithm based on artificial immunity of a bidding single agent provided by the present invention
  • FIG. 6 is a form of bid feedback provided by an exemplary embodiment of the present invention.
  • Figure 8 is an example of a microgrid structure
  • Figure 9 is a graph showing the trend of the affinity of antibody collection during an agent bidding process
  • Figure 10 is a demonstration of the clearance price of the microgrid for multiple coordination during the bidding period
  • Figure 11 shows the evolutionary algebraic distribution of the auction unit Agent for 5000 times
  • Figure 12 shows the coordinated evolutionary algebraic distribution of the microgrid solved 100 times to reach the Nash equilibrium.
  • the example provides a micro-grid optimization operation method of multi-agent structure, see FIG. 1 to FIG. 12, which are described in detail below:
  • micro-grid optimization architecture based on distributed power bidding in the embodiment is shown in Figure 1. It mainly involves three types of agents: micro-grid optimization control, bidding management, and bidding unit, which have the following characteristics.
  • Microgrid optimization control agent The agent sends the current microgrid operation status to the bid management agent, and receives the power generation and load energy balance information returned by the bid management agent.
  • the microgrid optimization control agent also includes monitoring the voltage, power quality and frequency of the microgrid.
  • the micro-grid optimization control agent under special circumstances (such as an accident in the large power grid, the micro-grid is switched from grid-connected to independent operation), the direct control method is used to carry out the load in the micro-grid and the distributed power generation output management. At this time, the bid management is ignored. Agent returns information.
  • Bidding management agent This agent is a function of assisting the micro-grid optimization control agent to complete the economic distribution of power generation in the micro-grid, and represents the public network purchasing/selling price of all loads and DER unit agents in the micro-grid on the main network. It also manages the bidding information of the bidding unit Agents in the microgrid. The bidding interval is consistent with the bidding time of the connected grid, usually in a period of 15 minutes.
  • the main functions of the Agent include: obtaining the bidding price of each DER, judging whether the negotiation is completed, the bidding information is filed, and the unit operation plan is obtained.
  • the bidding unit Agent mainly the DER unit, and may include the load unit Agent under certain circumstances.
  • the DER unit Agent is used to provide power to the energy source unit of the microgrid, such as a typical distributed energy source such as photovoltaic cells, wind turbines, fuel cells, micro turbines, and the like.
  • the DER unit agent submits a power generation bid request to the bid management agent based on local measurement information and communication information with other agents, and the time is 15 minutes.
  • the load unit is mainly for loads such as residential, commercial or industrial requirements in the microgrid, like the DER unit, negative
  • the unit can submit the expected amount of electricity and the highest and lowest price it is willing to pay.
  • the bid time is also 15 minutes, but does not participate in the auction.
  • Energy storage equipment can be regarded as power supply or load equipment. When discharging, it can be quoted according to DER. When charging, bid according to load.
  • the functions of the agent include: forming an antigen based on the published information, obtaining a bidding strategy through an artificial intelligence method, and submitting a bidding strategy.
  • the solution and implementation of the above bidding function of the microgrid is carried out by relying on the co-evolution of MAS and the adaptability and powerful search capability of the single control algorithm.
  • the general technology of the MAS system can be adopted, and the micro-grid bidding management agent collects the bidding information of each bidding agent through the inquiry method, and issues information through the blackboard system.
  • micro-grid bidding optimization operation control is: based on the research results of the existing MAS microgrid control system, the bidding agent function is added to the system, and transmitted to the control agent to apply the bidding result to guide the optimization operation.
  • the microgrid includes independent and grid-connected. In the grid-connected microgrid mode, the implementation time of the microgrid optimization operation is shown in Figure 2.
  • the main processes include the following:
  • the next period (1) connected to the main network to obtain micro grid sold to a user and the distributed power P buy electricity tariff repurchase P sell.
  • each DER in the micro-grid considers its own economic benefits and conducts optimization calculation based on bidding for microgrid.
  • the distributed power supply of the microgrid adopts the unit bidding agent to realize its own quotation function; the bid management agent assists in the negotiation and optimization of each bidding unit in the micro grid.
  • the bidding management agent sends the bidding result of each agent to the running agent, and the micro grid records the result based on the bidding negotiation and guides the microgrid optimization operation.
  • the key of this technology is the adaptive immune algorithm of single agent, and the realization of cooperative evolution of each agent based on this.
  • the micro-grid bidding feedback and co-evolutionary scheme are shown in Figure 3.
  • the technical idea is: the bidding unit DER is regarded as an agent with perceptual ability and rational thinking.
  • the agent interacts with the submitted bidding information (antibody representation).
  • the grid uses the bidding feedback provided by the bid management agent to realize the power supply and load in the micro grid and the power supply negotiation between the power sources.
  • An artificial immune solution process of the bidding unit Agent is a bidding process.
  • the antigenic form used by the bidding unit during the bidding process is shown in Figure 4.
  • the DER bidding unit first uses the self-environment and main grid price information as the antigen to perform artificial immune antibody solution to obtain the bid; then the bid management agent collects the bidding information and publishes it to each bidding unit agent in the form of bidding feedback.
  • the bidding unit agent generates a new antigen based on the feedback information and recalculates the bidding strategy. Through multiple rounds of feedback and negotiation, the final form of quotation is obtained.
  • the number of feedbacks allowed in the above bidding process determines the number of co-evolution of artificial immunity; the bidding of the microgrid The feedback determines the antigenic form of the artificial immune algorithm.
  • Different types of bidding models will vary.
  • the bidding feedback content received by the bidding unit may be the clearance price p cl or the bidding information of each bidding unit.
  • the feedback information p cl needs to be analyzed and analyzed by the bidding management agent.
  • the agent fails, the entire bidding function may collapse.
  • the bidding feedback is the bidding information of each agent, the DER bidding unit needs a larger amount of calculation.
  • the example uses the microgrid bidding process to allow multiple feedbacks (k is an integer greater than one), and the feedback is the quote for each bidding unit.
  • Other types of market bidding environments can use the same method, the specific antigen form and the maximum allowable feedback times of the bidding process (allowing the number of negotiations), and have no effect on the application of the method.
  • the auction feedback information in the method is the only way and content of the interaction between the agents.
  • the bid information of each bidding unit agent is used.
  • the initial value of the bidding information may be a real operating cost reflection, or may be a bidding information formulated through a strategy.
  • the optimal value obtained based on the artificial immune algorithm under the guidance of the set affinity is dynamic.
  • the bidding unit agent is based on the artificial immune algorithm flow, as shown in FIG. 5, and the following steps are specifically implemented.
  • the distributed power source owner determines the form of the offer and the objective function of the own bid based on the bidding model of the microgrid.
  • each DER bidding unit in the microgrid can use the quotation form of reporting four parameters [b i , c i , Q imin , Q imax ], and the price function and power limit form are:
  • B i (Q i ) is the bid price of the DER based on the generating capacity
  • b i and c i are the coefficients of the online bid of the i-th DER, and the bid directly determines whether the DER can win the bid and the amount of power generated by the bid.
  • Q imin and Q imax describe the maximum and minimum power generation during this time period, determined by distributed power unit characteristics and environmental factors.
  • the objective function ⁇ i (b i , c i ) of the distributed power bid is determined by the bidding method adopted by each distributed power source.
  • the bidding unit agent determines the antigenic representation of the artificial immune algorithm based on the characteristics of the distributed power source device.
  • the antigen model of each DER bidding unit may include three parts, (a) the environmental information of the bidding equipment itself, which is private information, and different equipments pay different attention, such as the micro-combustion machine pays attention to the natural gas price, and the solar power intensity in the time period of photovoltaic power generation attention The fan pays attention to the wind intensity during the time period. (b) During the optimized operation period, the price of electricity purchased by the microgrid from the main network and the price of electricity sold to the main network; (c) the quotation of other bidding unit agents, including the bidding coefficients b i and c of the distributed power source i , and the maximum and minimum power generation capacity.
  • the submitted bid feedback form is shown in Figure 6. This part of the information can be obtained from the bid management agent or through the information exchange between the agents.
  • the antibody represents the number of elements in the set n is the number of agents of the auction DER.
  • the environmental factors (antigen part 1) and their own technical parameters of each bidding DER determine the power generation cost, and the time period Q imin and Q imax ;
  • the market purchase price P buy and P sell (antigen part 2) determine the micro grid The upper and lower limits of the internal clearing price p cl ;
  • the bidding unit Agent's own quotation and bidding feedback information (antigen part 3) determine the clearing price P cl value of the micro grid, which is the core of the micro grid bidding, which is related to The interests of the various participants in the microgrid.
  • the interaction of information between agents is realized through the submission of bids, and is announced to other bidding agents in the form of bid feedback.
  • the bidding unit agent determines the antibody of the artificial immune algorithm and the encoding and decoding of the antibody based on the quotation form of the distributed power source.
  • the operation process of the artificial immune algorithm is not directly operated on the actual decision variable (power generation quotation) of the problem solved, but the cloning, mutation and selection of the individual codes representing the feasible solution, and the optimization is achieved by this immune method.
  • the bidding form of the bidding DER contains four parameters, [b i , c i , Q imin , Q imax ], and the antibody of the artificial immune algorithm forms a spatial correspondence relationship with the above four parameters through encoding and decoding.
  • an antibody as described includes only two b i and c i gene elements
  • the coding of the antibody is also two genetic elements for b i and c i .
  • the antibody coding of the exemplary example adopts the Gray code coding method, and its main advantage is that the change of adjacent values conforms to the principle of minimum character set coding, which is convenient for improving the local search capability of the AIS method.
  • the bidding unit agent determines the affinity formula of the artificial immune algorithm based on the objective function of the distributed power supply itself bidding.
  • the affinity of the antigen and the antibody reflects the bidding target and the profit of the bidding unit, and the affinity of the antibody and the antigen determines the pros and cons of the antibody, and the calculation of the affinity is determined by the bidding method adopted by each distributed power source.
  • the Agent is based on antigen and matures by increasing the affinity.
  • an antibody Ab can be formed and corresponds to a gain ⁇ i (b i , c i ), and the size of the gain determines the affinity of the antibody.
  • the affinity is considered to be the highest.
  • the bidding unit Agent Based on the antigen expression form of step 402, the bidding unit Agent processes the collected information to form an artificially immunized environmental antigen.
  • the auction feedback information in the method is the only way and content of the interaction between the agents.
  • the bid information of each bidding unit agent is used.
  • the initial value of the bidding information may be a real operating cost reflection, or may be a bidding information formulated through a strategy.
  • the optimal value obtained based on the artificial immune algorithm under the guidance of the set affinity is dynamic.
  • the unit bidding agent is based on the quotation environment antigen, and is solved by an artificial immune algorithm to obtain an antibody that conforms to the interests of the distributed power source, and the antibody is decoded to obtain the bidding scheme of the distributed power source, and submitted to the bid management agent.
  • each bidding process is an immunization process.
  • the bidding process can use various intelligent algorithms with adaptive, including but not limited to artificial immune algorithms, genetic algorithms and other intelligent algorithms.
  • the bidding unit Agent cooperatively evolves and performs steps 405 and 406.
  • the bidding management agent assists the bidding unit agent in the micro-grid to achieve negotiation and optimization by collecting and sending bidding information.
  • the specific implementation steps include:
  • 502 Wait and accept bid information submitted by each bidding unit agent in the micro grid.
  • the judgment conditions include two types: (a) reaching the specified number of consultations; (b) the clearance power price in the micro grid does not change after the negotiation.
  • microgrid negotiation bid allowed to be based on the number of max bid market rules microgrid determined, in the case of allowing multiple negotiations, preferably an integer of 3 to 10.
  • step 505 If it is judged that the negotiation is completed, the process proceeds to step 501, and the bidding unit agent in the micro power grid is negotiated by repeating the bidding manner.
  • the application of this method is mainly considered.
  • the microgrid adopts a centralized and distributed hybrid control method, and adopts a release form with higher efficiency. If there is an information interaction channel between the bidding units, the participation of the bid management agent is not required, and the information interaction between the bidding unit agents can be realized.
  • the artificial immune algorithm is a parallel optimization calculation process based on the biological immune mechanism.
  • the above-mentioned various bidding unit agents only search for antigens in their own bidding space, and search through the antibody set and the gene pool, the process does not need central control. Assist, complete in parallel.
  • the antigen of each bidding unit agent contains bidding feedback information (such as other antibody bidding information content or clearance price cl ) issued from the bidding management agent, so the antigen content is global.
  • the micro-grid bidding process of AIS system based on co-evolution is shown in Fig. 7.
  • the application demonstration is carried out for the commercial microgrid project shown in Figure 8 for the system structure.
  • the microgrid is divided into three feeders, including two micro-combustion engines, two diesel engines, and four different types of distributed power sources such as solar photovoltaic solar energy.
  • the reactive part does not need to be considered.
  • the cost factors of each distributed power supply and the upper and lower limits of the power generation are shown in Table 1.
  • the price of microgrid purchases and sales from the main network is 0.022$/kWh and 0.03$/kWh, respectively. Since the time period of each stage is fixed (15 min), the power in the micro grid can be used instead of the power. The active power demand of the three loads during this period is 130 kW, and the reactive power demand and influence are not considered in the micro grid.
  • the various bidding unit agents in the micro-grid will achieve coordinated evolution. Since each part is based on its own private information, it will be based on other agent quotations, and the bid price will converge quickly to reach the Nash equilibrium state.
  • the bid price convergence in the microgrid is shown in Figure 10.
  • the microgrid agent can achieve the minimum power generation and transmission cost of the microgrid while maintaining the maximum utilization efficiency of photovoltaic and wind power generation.
  • the bidding results are shown in Table 2.
  • the artificial immune algorithm is a random algorithm.
  • the running time of the system may be different.
  • This method performs statistical analysis on the expected algebra of the optimal solution for a single agent.
  • the system runs 5000 times on the micro-combustion engine Agent bidding process.
  • the statistical results of the adaptive bidding scheme are shown in Fig. 11.
  • the horizontal axis represents evolutionary algebra, and the vertical axis represents that the global optimal solution can be found after running to the corresponding evolutionary algebra. Number of times (frequency). It can be seen from the figure that the global optimal solution is found with high probability within 30 generations, and the average algebra of the optimal solution is found to be 13.5 generations.
  • Figure 12 is a case where 100 microgrid co-evolutions are obtained to obtain an optimal result.
  • the maximum expected number of Nash equilibrium prices based on bidding negotiation is 5 times. In order to save runtime, fewer feedbacks and negotiated times can be defined to complete the bidding within the microgrid.
  • the invention runs on the basis of the antibody gene pool, and ensures rapid convergence and global optimization of the artificial immune bidding scheme.
  • the bidding process is an iterative process. In practice, the anomalous events are far less than the normal events.
  • the new antibodies generated during the initial response process have memory.
  • the second response mechanism is used to quickly obtain the optimal bidding scheme. .
  • each agent is running in parallel calculation, and the overall efficiency is high.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Economics (AREA)
  • Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Power Engineering (AREA)
  • General Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Marketing (AREA)
  • Human Resources & Organizations (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • Tourism & Hospitality (AREA)
  • Water Supply & Treatment (AREA)
  • Public Health (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Development Economics (AREA)
  • Technology Law (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Software Systems (AREA)
  • Automation & Control Theory (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

本发明公开了一种多智能体结构的微电网优化运行方法，所述方法包括以下步骤：微电网获得连接主网下一时段出售给用户的电价Pbuy和分布式电源回购的电价Psell；进行基于竞价的微电网优化运行计算；微电网分布式电源采用单元竞价实现报价功能；竞价管理智能体协助完成微电网内各单元的协商和优化；竞价管理智能体将各单元发电量竞价结果发送给微电网控制系统，微电网控制系统基于竞价协商后的结果进行微电网的优化运行。本发明能够解决当微电网内设备产权不统一，或涉及公共微电网时，现有运行方式无法进行微电网内部利益分配的问题，有助于微电网商业化运营的实现。

Description

一种多智能体结构的微电网优化运行方法 技术领域
本发明涉及电力系统领域，尤其涉及一种多智能体结构的微电网优化运行方法。
背景技术
微电网作为分布式电源(Distributed Energy Resource，DER)的重要组织方式，得到了极大的关注，其中微电网的优化运行更成为研究的热点[1]。多智能体系统(Multi-Agent Systems，MAS)适宜于解决复杂的、开放的分布式问题，近年来广泛应用于微电网的运行控制，其相关技术也受到关注，并取得一定的成果。
目前采用MAS结构的微电网优化运行多以网设备发电、运行维护成本或与主网互购电能费用为基础的最小成本为目标[2]，内部基于自愿协作方式，通过各智能体(Agent)的完全合作方式进行分工协作。该类运行方式在微电网内设备产权单一的情况下较为简单、实用。随着微电网市场化的发展，当微电网内出现不同投资主体时，如多家酒店组成的微电网，该类运行方式无法解决微电网内部利益分配的问题。在开放的市场环境中，微电网内各DER允许有自己的目标和利益，可以采取有意图的方式参与微电网的运行，此时基于自愿协作仅仅是Agent间相互作用的一种特殊情形。
通过竞价的市场机制引导微电网内部各主体参与有效竞争，是未来微电网商业化运营的一种重要方式[3]。竞价方法配合微电网的控制方式，可以通过发电竞价的制定实现分布式电源的分散决策，同时也可以运用价格协调管理达到集中控制。由于分布式电源及微电网自身的特性，现有常规电力市场研究成果并不能简单的移植到微电网中，主要体现在以下两个方面：
(1)常规电力市场发电竞价通常作为独立的系统，发电商根据各自竞价策略和运行情况形成报价，汇集到电力市场管理机构竞价，并形成发电计划，这些操作与电力系统的运行、控制分开完成；而在基于MAS结构的微电网中，系统的运行决策是通过各种Agent完成的，竞价结果将直接传递给微电网运行控制Agent执行，因此竞价功能需要融合在MAS运行控制的整体结构中，而不能单独设计。
(2)发电竞价受市场和技术等多种因素影响，是个复杂的技术经济问题。分布式电源，如风电和太阳能光伏发电，具有间歇性和不可控性，运行环境复杂，通过实际观测、或对实际系统的分析和仿真，很难获取包容所有模式下的竞价方案样本。目前常规竞价算法，如“强 化学习”、“重复博弈”等算法，还无法确保对未知环境下的竞价获得较满意的结果。微电网中由于竞价结果直接指导运行控制，面上表现“理性”的竞价Agent在突发事件面前可能表现得很“笨拙”，这是难以接受的[4]。
基于上述问题，通过在微电网已有运行控制的基础上增添分布式的竞价功能，通过市场手段反映各微电网参与者的利益诉求，并指导微电网优化运行；竞价过程中利用人工免疫系统(Artificial Immune systems，AIS)的自适应和缺陷容忍能力，处理间歇性电源所带来的不确定性问题；并通过该技术的协同进化过程，提高整个微电网MAS的协调性。
发明内容
本发明提供了一种多智能体结构的微电网优化运行方法，本发明能够解决当微电网内设备产权不统一，或涉及公共微电网时，现有运行方式无法实现微电网内部利益分配的问题。该方法通过竞价的市场机制引导微电网优化运行，具体过程见下文描述：
一种多智能体结构的微电网优化运行方法，所述方法包括以下步骤：
(1)微电网获得连接主网下一时段出售给用户的电价Pbuy和分布式电源回购的电价Psell；
(2)进行基于竞价的微电网优化运行计算；微电网分布式电源采用单元竞价实现报价功能；竞价管理智能体协助完成微电网内各单元的协商和优化；
(3)竞价管理智能体将各单元发电量竞价结果发送给微电网控制系统，微电网控制系统基于竞价协商后的结果进行微电网的优化运行。
所述微电网分布式电源采用单元竞价实现报价功能的步骤具体为：
基于分布式电源自身竞价的目标函数，确定人工免疫算法的亲和度公式，形成人工免疫的报价环境抗原；
通过对人工免疫算法进行求解，获得符合分布式电源自身利益的抗体，将抗体解码获得分布式电源的竞价方案，并提交给竞价管理智能体。
所述竞价管理智能体协助完成微电网内各单元的协商和优化的步骤具体为：
判断微电网内分布式电源和负荷之间是否达成平衡；
如果是，则备案各竞价单元的竞价情况，并获得机组运行方案，提交给微电网控制系统；
如果否，通过多次竞价方式使微电网内分布式电源和负荷之间达成平衡。
所述基于分布式电源自身竞价的目标函数，确定人工免疫算法的亲和度公式，形成人工免疫的报价环境抗原具体步骤为：
(1)分布式电源所有者基于微电网的竞价模型，确定报价形式，以及自身竞价的目标函 数；
(2)竞价单元智能体基于分布式电源自身特性，确定人工免疫算法的抗原表现形式；
(3)竞价单元智能体基于分布式电源的报价形式，确定人工免疫算法的抗体，以及抗体的编码和解码公式；
(4)竞价单元智能体基于分布式电源自身竞价的目标函数，确定人工免疫算法的亲和度公式；
(5)竞价单元智能体对收集到的信息进行加工，形成人工免疫的报价环境抗原。
其中，人工免疫算法的抗原表现形式包括以下三部分信息：
(a)竞价设备自身的环境信息；
(b)优化运行时间内，微电网从连接主网的购电价格以及出售给主网的售电价格；
(c)其他单元竞价智能体t的报价信息。
判断条件为：达到规定的次数，和/或微电网内的清仓电价不再变化。
本发明提供的技术方案的有益效果是：
(1)未来电力市场及微电网的发展，微电网组成可能属于不同业主，基于竞价等经济手段进行优化运行与控制不可避免。本发明能够解决当微电网内设备产权不统一，或涉及公共微电网时，现有运行方式无法进行微电网内部利益分配的问题，有助于微电网商业化运营的实现。
(2)本发明采用MAS的竞价优化运行与控制，通过市场手段反映各微电网参与者的利益诉求，并利用竞价结果指导微电网优化，符合微电网分层、分布式控制的特点。对于分布式电源或可控负荷单元来说，生产制造的控制器中依据MAS标准嵌入该发明算法的可编程智能体，即可实现“即插即用”和直接参与系统竞价优化运行。
附图说明
图1为本发明提供的基于竞价的微电网优化控制框架；
图2为微电网优化运行实施时间过程示意图；
图3为本发明提供的微电网竞价反馈和协同进化方案；
图4为本发明提供的竞价单元人工免疫算法的抗原形式；
图5为本发明提供的竞价单体Agent基于人工免疫的优化算法流程；
图6为本发明示范例提供的竞价反馈形式；
图7为实施例基于协同进化的AIS系统模型框架；
图8为示范例的微电网结构；
图9为示范例某Agent竞价过程中抗体集合亲和度的变化趋势；
图10为示范例竞价时段内微电网多次协调得到清仓价格；
图11为竞价单元Agent求解5000次的进化代数分布；
图12为微电网求解100次达到纳什均衡的协调进化代数分布。
具体实施方式
为使本发明的目的、技术方案和优点更加清楚，下面对本发明实施方式作进一步地详细描述。
为了满足商业化微电网运行过程中属于不同业主的微网组成单元各自的利益诉求，同时考虑微电网分层、分布式控制，含有多种类型间歇性不可控分布式电源的特点，本发明实施例提供了一种多智能体结构的微电网优化运行方法，参见图1至图12，详见下文描述：
一、微电网优化运行的架构及Agent设置；
实施例中基于分布式电源竞价的微电网优化架构如图1所示，主要涉及微电网优化控制、竞价管理，以及竞价单元三种类型Agent，分别具有如下特征。
(1)微电网优化控制Agent：该Agent给竞价管理Agent发送当前微电网的运行情况，并接收竞价管理Agent返回的发电、负荷能量平衡信息。微电网优化控制Agent除了进行微电网能量管理的控制，还包括对微电网内电压、电能质量、频率等进行监控。微电网优化控制Agent在特殊情况下(如大电网发生事故，微电网由并网转为独立运行)，采用直接控制方式进行微电网内的负荷和分布式电源发电出力管理，此时忽略竞价管理Agent返回信息。
(2)竞价管理Agent：该Agent是辅助微电网优化控制Agent完成微电网内发电经济分配方面的功能，并代表主网对微电网内所有负荷和DER单元Agent公开购电/售电价格，同时也管理微电网内各竞价单元Agent的招标投标信息。竞价间隔和所接入电网竞价时间保持一致，通常为15分钟一个阶段。该Agent的主要功能包括：获得各DER的发电竞价，判断是否协商完毕，竞价信息备案，获得机组运行方案。
(3)竞价单元Agent：主要为DER单元，一定情况下可包括负荷单元Agent。DER单元Agent用于提供电力到微电网的能量源单元，如光伏电池，风力涡轮机，燃料电池，微涡轮机等典型的分布式能量源。DER单元Agent基于本地测量信息以及与其他Agent的通信信息，而提交发电竞价请求给竞价管理Agent，时间是15分钟一次。
负荷单元主要针对微电网内的如住宅，商业或工业等需求的负载，如DER单元一样，负 荷单元可以提交期望获得的电量及其愿意支付的最高和最低价格，该投标时间也是15分钟一次，但不参与竞价。储能设备可以视为电源或负荷设备，当放电时候，可以按照DER报价，当充电时，按照负荷投标。该Agent的功能包括：基于发布的信息形成抗原，通过人工智能方法获得竞价策略，提交竞价策略。
实施例中微电网上述竞价功能的求解与实现依靠MAS的协同进化和单体控制算法的自适应性和强大搜索能力进行的。通信方面，可采用MAS系统的通用技术，微电网竞价管理Agent通过询问方式收集各竞价Agent的竞价信息，并通过黑板系统发布信息。
二、微电网优化运行实施流程
微电网竞价优化运行控制思想为：以现有MAS微电网控制系统研究成果为基础，在该系统上增添竞价Agent功能，并传输给控制Agent应用竞价结果指导优化运行。微电网包括独立型和并网型，在并网型微电网模式下，实施例进行微电网优化运行实施时间过程如图2所示。
主要流程包括如下：
(1)微电网获得连接主网下一时段出售给用户的电价Pbuy和分布式电源回购的电价Psell。
(2)微电网内各DER考虑自身的经济效益，进行基于竞价的微电网优化运行计算。实施过程微电网分布式电源采用单元竞价Agent实现自身的报价功能；竞价管理Agent协助完成微电网内各竞价单元的协商和优化。
(3)竞价管理Agent将各Agent的发电量竞价结果发送给运行Agent，微电网基于竞价协商后的结果进行备案，并指导微电网优化运行。
三、微电网竞价过程中的竞价反馈和协同进化方案
基于人工免疫的微电网分布式电源的竞价算法，该技术的关键是单体Agent的自适应免疫算法，以及在此基础上所实现的各Agent的协同进化的实现。微电网竞价反馈和协同进化方案如图3所示，其技术思想为：将竞价单元DER视为具有感知能力和理性思维的Agent，Agent之间通过提交的竞价信息(抗体代表)发生交互，微电网通过竞价管理Agent提供的竞价反馈实现微电网内电源与负荷，以及电源之间的供电协商。
竞价单元Agent的一次人工免疫求解过程即是一次竞价过程。竞价过程中竞价单元采用的抗原形式如图4所示。在一个竞价时段内，DER竞价单元首先以自身环境和主网电价信息作为抗原，进行人工免疫抗体求解获得竞价；之后竞价管理Agent收集竞价信息，并以竞价反馈的形式公布给各个经竞价单元Agent，竞价单元Agent基于反馈信息生成新的抗原，并重新计算获得竞价策略。通过多轮的反馈及协商，得到最终的报价形式。
其中，上述竞价过程中允许的反馈次数决定了人工免疫的协同进化次数；微电网的竞价 反馈内容决定了人工免疫算法的抗原形式。不同类型的竞价模式会有所不同。
(1)若以k表示微电网允许各竞价单元反馈和修改竞价的最大次数，当微电网不允许反馈，则有k＝0，目前大电网主要采用此种形式；当微电网允许进行一次反馈，此时k＝1；当微电网允许通过多轮反馈而达到网内电能供需的协商，则k不受限制时。通常情况下，经过少数几次反馈即可达到纳什均衡。
(2)当允许进行有反馈的竞价时，竞价单元收到的竞价反馈内容可以是清仓价格pcl，也可以是各个竞价单元的竞价信息。不同之处在于反馈信息pcl需要由竞价管理Agent分析获得，当该Agent出现故障时，整个竞价功能可能崩溃；而竞价反馈为各Agent的竞价信息时，DER竞价单元需要更大的运算量分析报价情况并给出自身最优的竞价策略。但后者既可通过竞价管理Agent的收集和发布获得，也可以通过各个竞价单元Agent之间的信息交互获得，更适应未来分布式的微电网控制方式。
为具有普遍性，示范例采用微电网竞价过程允许多次反馈(k为大于1的整数)，反馈为各个竞价单元的报价。其他类型的市场竞价环境可以采用相同方法，具体抗原形式及竞价过程的最大允许反馈次数(允许协商次数)，对本方法的应用不产生影响。
其中，本方法中竞价反馈信息是Agent之间交互的唯一途径和内容，本实施例采用各竞价单元Agent的竞价信息。该竞价信息初始值可以是自身真实的运行成本反映，也可以是通过策略制定的竞价信息。但在竞价过程中，均是在设定的亲和度指导下，基于人工免疫算法获得的最优值。在同一运行控制时间内，各Agent提交的竞价随着迭代而不断发生变化，因此抗原是动态的。
四、实施例中竞价单元智能体基于人工免疫的算法流程，如图5所示，具体采用如下步骤实现。
401：分布式电源所有者基于微电网的竞价模型，确定报价形式，以及自身竞价的目标函数。
其中，微电网内各DER竞价单元可采用上报4个参数[bi,ci,Qimin,Qimax]的报价形式，价格函数和功率限制形式为：
Bi(Qi)＝bi+ciQi，                 (1)
s.t.Qimin≤Qi≤Qimax                  (2)
式中，Bi(Qi)为DER基于发电容量的竞价价格，bi和ci为第i台DER的上网竞价的系数，该竞价将直接决定DER能否中标，以及中标的发电量。Qimin和Qimax描述在该时段的最大和最小发电量，由分布式电源机组特性和环境因素决定。
分布式电源竞价的目标函数πi(bi,ci)由各分布式电源采用的竞价方法决定。
402：竞价单元Agent基于分布式电源自身设备特性，确定人工免疫算法的抗原表现形式。
各DER竞价单元的抗原模型可包括三部分，(a)竞价设备自身的环境信息，该部分属于私有信息，不同设备关注不同，如微燃机关注天然气价格，光伏发电关注时间段内的日照强度，风机关注时间段内的风强度。(b)优化运行时间段内，微电网从连接主网的购电价格以及出售给主网的售电价格；(c)其他竞价单元Agent的报价，包括分布式电源的竞价系数bi和ci，以及最大最小发电容量。提交的竞价反馈形式如图6所示，该部分信息可从竞价管理Agent获得的，也可通过Agent之间的信息交互获得。当仅允许DER参与竞价时，抗体代表集合内元素个数n即为竞价DER的Agent数。
其中，各竞价DER的环境因素(抗原部分1)和自身技术参数，决定了发电成本，以及该时段Qimin和Qimax；市场购售电价Pbuy和Psell(抗原部分2)决定了微电网内结清电价pcl的上下限范围；竞价单元Agent的自身报价和竞价反馈信息(抗原部分3)，决定了微电网的结清电价Pcl值，该值是微电网竞价的核心，关系到微电网内各参与者的利益。Agent之间信息的交互是通过竞价的提交实现，并以竞价反馈的方式公布给其他竞价Agent。
403：竞价单元Agent基于分布式电源的报价形式，确定人工免疫算法的抗体，以及抗体的编码和解码。
人工免疫算法的运行过程，不是对所求解问题的实际决策变量(发电报价)直接进行操作，而是对表示可行解的个体编码进行克隆、变异和选择等运算，通过这种免疫方式达到优化的目的。竞价DER的报价形式包含四个参数，[bi,ci,Qimin,Qimax]，人工免疫算法的抗体通过编码和解码与上述四个参数形成空间对应关系。
由于Qimin和Qimax与所在阶段的环境和技术因素有关，实施例可以选择Qimin和Qimax不参与竞价策略，此种情况下所描述的抗体仅包括bi和ci两个基因元素，抗体的编码也是对bi和ci两个基因元素。示范例的抗体编码采用格雷码编码方式，其主要优点是相邻值的变化符合最小字符集编码原则，便于提高AIS方法的局部搜索能力。
404：竞价单元Agent基于分布式电源自身竞价的目标函数，确定人工免疫算法的亲和度公式。
抗原和抗体的亲和度体现竞价单元的竞价目标和收益，同时抗体和抗原的亲和度决定该抗体的优劣，亲和度的计算由各分布式电源采用的竞价方法决定。
Agent基于抗原，通过亲和度的增加使抗体进化成熟。对于Agent任意解空间里的一个(bi,ci)，都能够组成一个抗体Ab，并对应一个收益πi(bi,ci)，此收益的大小决定了抗体的亲和度。 在一定的环境下(抗原)，竞价策略(抗体)能够获得最大的收益(目标函数)时，认为亲和度最高。
405：基于步骤402的抗原表现形式，竞价单元Agent对收集到的信息进行加工，形成人工免疫的环境抗原。
其中，本方法中竞价反馈信息是Agent之间交互的唯一途径和内容，本实施例采用各竞价单元Agent的竞价信息。该竞价信息初始值可以是自身真实的运行成本反映，也可以是通过策略制定的竞价信息。但在竞价过程中，均是在设定的亲和度指导下，基于人工免疫算法获得的最优值。在同一运行控制时间内，各Agent提交的竞价随着迭代而不断发生变化，因此抗原是动态的。
406：单元竞价Agent基于报价环境抗原，通过人工免疫算法进行求解，获得符合分布式电源自身利益的抗体，将抗体解码获得该分布式电源的竞价方案，并提交给竞价管理Agent。
其中，每次竞价过程，即是一次免疫过程。竞价求解过程可以采用各种具有自适应的智能算法，包括但不限于人工免疫算法，遗传算法等智能算法。
微电网优化运行过程中，竞价单元Agent协同进化并执行步骤405和406。
五、实施过程中竞价管理Agent，通过收集和发送竞价信息等操作，协助微电网内竞价单元Agent实现协商和优化，具体实现步骤包括：
501：将收集到的抗体代表发给各竞价Agent。
502：等待并接受微电网内各竞价单元Agent提交的竞价信息。
503：基于收集到的各竞价信息，判断微电网内分布式电源和负荷之间是否完成协商，达成平衡。判断条件包括两种：(a)达到规定的协商次数；(b)协商后微电网内的清仓电价不再变化。
其中，微电网允许竞价的协商次数maxbid需根据微电网的市场规则确定，对于允许多次协商的情况，可取3～10之间某整数。
504：如果判断完成协商为是，则备案各竞价单元的竞价情况，并获得机组运行方案，提交给微电网运行管理Agent。
505：如果判断完成协商为否，转步骤501，通过重复竞价方式使微电网内各竞价单元Agent实现协商。应用该方法主要是考虑的目前微电网采用的是集中和分布式混合控制方法，采用发布形式具有较高效率。若各竞价单元之间有信息交互通道，则不需要竞价管理Agent的参与，可以通过各竞价单元Agent之间信息交互而实现。
其中，人工免疫算法是建立在生物免疫机理的一种并行优化计算过程，上述各个竞价单 元Agent只在自己的竞价空间中针对抗原，通过抗体集和基因库进行搜索，该过程不需要中央控制的协助，并行完成。但各竞价单元Agent的抗原中包含从竞价管理Agent发布的竞价反馈信息(如其他的抗体竞价信息内容或清仓价格cl)，因此抗原内容具有全局性。基于协同进化的AIS系统微电网竞价过程如图7所示。
算例和分析
针对系统结构如图8所示的商用微电网项目进行应用示范。该微电网分为三个馈线，包括2台微燃机、2台柴油机，以及4个太阳能光伏太阳能等不同类型的分布式电源。无功部分不需要考虑，各分布式电源的成本因素及发电功率上下限如表1所示。
表1分布式电源的成本因素及发电功率上下限
Figure PCTCN2015077471-appb-000001
若某高峰时刻，微电网从主网购电和售电的价格分别为0.022$/kWh和0.03$/kWh。由于各个阶段的时间段固定(15min)，因此可用微电网内的功率来代替电量，三个负荷在该时段的有功功率需求量为130kW，微电网内不考虑无功需求及影响。算法中主要运行参数默认值为：抗体集规模m＝8，抗体克隆规模sizeclone＝4，变异模式门槛值P＝0.1，允许竞价协商最大次数maxbid＝5，抗体进化允许最大代数MaxGen＝50。
某Agent竞价过程中抗体集合亲和度的变化趋势如图9所示。图中横轴为抗体进化迭代代数，纵轴为DER获得的收益(适应度)。图中的两条线分别为抗体集合的亲和度最大值和平均值。由图可以看出，对于每个Agent，自身抗体(竞价策略)对于抗原(环境因素和其他抗体代表集合)的亲和度虽然有波动情况，但总体是上升的。
通过抗体代表的提呈，微电网内各个竞价单元Agent会实现协同进化。由于各个部分基于自身的私有信息进行的竞价，但会基于其他agent报价的情况进行，竞标价格很快就收敛，达到纳什均衡状态。微电网内竞标价格收敛情况如图10所示。
通过这种Agent竞价策略的发电调整后，微电网Agent可以在保持光伏和风力发电最大利用效率的情况下，达到微电网的发电、输电成本最小化。竞价结果如表2所示。
表2基于竞价的微电网发电运行结果
Figure PCTCN2015077471-appb-000002
注：获得收益仅包括电能的收益，供冷热的效益已经在DER报价中减去。
人工免疫算法是一种随机算法，系统的每次运行时间都有可能不一样，本方法对单个Agent可以获得最优解的期望代数进行了统计分析。系统对微燃机Agent竞价过程运行5000次，获得自适应竞价方案的统计结果如图11所示，其中横轴表示进化代数，纵轴表示在运行到相应的进化代数可以找到全局最优解的次数(频率)。由图可见，在30代内以很高概率找到全局最优解，发现最优解的平均代数为13.5代。
对于整个微电网需要多次的协同进化才能达到纳什均衡。图12是求解100次微电网协同进化获得最优结果的情况。由图可见，基于竞价的协商能够获得纳什均衡价格的最大期望次数为5次。为了节省运行时间，可以限定更少的反馈和协商次数完成微电网内的竞价。
其中，本发明在抗体基因库的基础上运行，保证了人工免疫竞价方案的快速收敛和全局最优。竞价过程是一个不断重复的过程，实际当中异常事件远远少于正常事件，初次应答过程中产生的新抗体具有记忆性，当相似问题再次出现时，采用二次应答机制快速获得最优竞价方案。同时各个Agent处于并列运行计算，整体效率较高。
参考文献：
[1]王成山，武震，李鹏.微电网关键技术研究[J].电工技术学报,2014,29(2):1-12.
[2]艾芊,章健.基于多代理系统的微电网竞价优化策略[J].电网技术，2010,34(2):46-51.
[3]孔祥玉，房大中.考虑网损的联营体日前发电竞价模型研究[J].武汉大学学报(工学版),2009,42(2):105-109.
[4]宋依群,高瞻.Agent技术在电力市场中的应用综述[J].电力系统及其自动化学报，2008,20(3)：111-116.
本领域技术人员可以理解附图只是一个优选实施例的示意图，上述本发明实施例序号仅仅为了描述，不代表实施例的优劣。
以上所述仅为本发明的较佳实施例，并不用以限制本发明，凡在本发明的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。

Claims (6)

  1. 一种多智能体结构的微电网优化运行方法，其特征在于，所述方法包括以下步骤：
    (1)微电网获得连接主网下一时段出售给用户的电价Pbuy和分布式电源回购的电价Psell；
    (2)进行基于竞价的微电网优化运行计算；微电网分布式电源采用单元竞价实现报价功能；竞价管理智能体协助完成微电网内各单元的协商和优化；
    (3)竞价管理智能体将各单元发电量竞价结果发送给微电网控制系统，微电网控制系统基于竞价协商后的结果进行微电网的优化运行。
  2. 根据权利要求1所述的一种多智能体结构的微电网优化运行方法，其特征在于，所述微电网分布式电源采用单元竞价实现报价功能的步骤具体为：
    基于分布式电源自身竞价的目标函数，确定人工免疫算法的亲和度公式，形成人工免疫的报价环境抗原；
    通过对人工免疫算法进行求解，获得符合分布式电源自身利益的抗体，将抗体解码获得分布式电源的竞价方案，并提交给竞价管理智能体。
  3. 根据权利要求1所述的一种多智能体结构的微电网优化运行方法，其特征在于，所述竞价管理智能体协助完成微电网内各单元的协商和优化的步骤具体为：
    判断微电网内分布式电源和负荷之间是否达成平衡；
    如果是，则备案各竞价单元的竞价情况，并获得机组运行方案，提交给微电网控制系统；
    如果否，通过多次竞价方式使微电网内分布式电源和负荷之间达成平衡。
  4. 根据权利要求2所述的一种多智能体结构的微电网优化运行方法，其特征在于，所述基于分布式电源自身竞价的目标函数，确定人工免疫算法的亲和度公式，形成人工免疫的报价环境抗原具体步骤为：
    (1)分布式电源所有者基于微电网的竞价模型，确定报价形式，以及自身竞价的目标函数；
    (2)竞价单元智能体基于分布式电源自身特性，确定人工免疫算法的抗原表现形式；
    (3)竞价单元智能体基于分布式电源的报价形式，确定人工免疫算法的抗体，以及抗体的编码和解码公式；
    (4)竞价单元智能体基于分布式电源自身竞价的目标函数，确定人工免疫算法的亲 和度公式；
    (5)竞价单元智能体对收集到的信息进行加工，形成人工免疫的报价环境抗原。
  5. 根据权利要求4所述的一种多智能体结构的微电网优化运行方法，其特征在于，人工免疫算法的抗原表现形式包括以下三部分信息：
    (a)竞价设备自身的环境信息；
    (b)优化运行时间内，微电网从连接主网的购电价格以及出售给主网的售电价格；
    (c)其他单元竞价智能体t的报价信息。
  6. 根据权利要求3所述的一种多智能体结构的微电网优化运行方法，其特征在于，判断条件为：
    达到规定的次数，和/或微电网内的清仓电价不再变化。
PCT/CN2015/077471 2014-11-19 2015-04-26 一种多智能体结构的微电网优化运行方法 WO2016078329A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US15/526,738 US20190108600A1 (en) 2014-11-19 2015-04-26 Optimized Operation Method in Micro-Grid with Multi Agent Structure

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201410668588.8 2014-11-19
CN201410668588.8A CN104392279B (zh) 2014-11-19 2014-11-19 一种多智能体结构的微电网优化运行方法

Publications (1)

Publication Number Publication Date
WO2016078329A1 true WO2016078329A1 (zh) 2016-05-26

Family

ID=52610179

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2015/077471 WO2016078329A1 (zh) 2014-11-19 2015-04-26 一种多智能体结构的微电网优化运行方法

Country Status (3)

Country Link
US (1) US20190108600A1 (zh)
CN (1) CN104392279B (zh)
WO (1) WO2016078329A1 (zh)

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108075568A (zh) * 2016-11-10 2018-05-25 中国电力科学研究院 一种电力系统电压危急状态的监控系统及其构建方法
CN110111135A (zh) * 2019-04-09 2019-08-09 广东电力交易中心有限责任公司 一种发电侧成员动态报价决策方法、装置及设备
CN112215687A (zh) * 2020-11-17 2021-01-12 东北大学 一种基于云边协同的家庭微网自主竞价系统及方法
CN113377073A (zh) * 2021-06-28 2021-09-10 西南交通大学 一种基于双层多智能体系统的柔性作业车间调度优化方法
CN113570221A (zh) * 2021-07-15 2021-10-29 国网浙江省电力有限公司经济技术研究院 基于动态演化视角的电网企业拓展综合能源服务市场的辅助决策方法
CN113705067A (zh) * 2021-10-29 2021-11-26 中国电力科学研究院有限公司 一种微网优化运行策略生成方法、系统、设备及存储介质
CN114675535A (zh) * 2022-03-07 2022-06-28 大连理工大学 一种基于强化学习的航空发动机过渡态寻优控制方法
CN114721409A (zh) * 2022-06-08 2022-07-08 山东大学 一种基于强化学习的水下航行器对接控制方法
CN116451880A (zh) * 2023-06-16 2023-07-18 华北电力大学 一种基于混合学习的分布式能源优化调度方法及装置
CN117154954A (zh) * 2023-10-31 2023-12-01 国网浙江省电力有限公司宁波供电公司 一种基于遗传算法的配电网馈线自动化终端的配置方法

Families Citing this family (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104392279B (zh) * 2014-11-19 2018-02-13 天津大学 一种多智能体结构的微电网优化运行方法
CN105186583B (zh) * 2015-10-22 2017-11-03 东北大学 基于多智能体建模的能量路由器及其能量调度方法
CN106451415B (zh) * 2016-09-07 2018-12-21 广东顺德中山大学卡内基梅隆大学国际联合研究院 一种基于分布式协调方法的微电网电力交易系统
US20180150857A1 (en) * 2016-11-30 2018-05-31 Electronics And Telecommunications Research Institute Apparatus and method for processing tender of aggregate power plant
CN106602555A (zh) * 2016-12-27 2017-04-26 上海中兴电力建设发展有限公司 具有分层分布式结构的能源互联网能量管理系统
CN106684869B (zh) * 2017-03-17 2019-05-28 燕山大学 一种考虑内外博弈的主动配电网故障恢复策略
CN107026448A (zh) * 2017-05-22 2017-08-08 上海海事大学 基于多智能体的混合动力船舶微电网系统
CN107958300B (zh) * 2017-08-21 2020-08-04 浙江工业大学 一种考虑互动响应的多微网互联运行协调调度优化方法
CN108199379A (zh) * 2018-01-17 2018-06-22 湖南世优电力科技股份有限公司 基于多智能体免疫算法的集中式电压控制装置及控制方法
CN108333940B (zh) * 2018-02-12 2021-08-06 北京建筑大学 一种优化指定控制器参数的方法及装置
CN108596692A (zh) * 2018-05-11 2018-09-28 南京邮电大学 基于多Agent技术的微电网电力市场竞价方法
CN110618690B (zh) * 2018-06-20 2022-06-07 陕西师范大学 一种基于免疫网络与市场机制的多uuv系统动态协商建模方法
CN110729722B (zh) * 2019-10-22 2023-11-07 东北大学 一种基于能源细胞架构的区域能源网优化调度方法
EP4050547A4 (en) * 2019-10-25 2023-11-15 Kyocera Corporation TENDER ASSISTANCE SYSTEM AND TENDER ASSISTANCE PROCESS
CN111489031B (zh) * 2020-04-09 2022-06-03 江苏方天电力技术有限公司 基于源荷演化博弈的综合能源系统中长期负荷预测系统及方法
US11695299B2 (en) * 2020-10-01 2023-07-04 Tianjin University Quick-response voltage control method of distribution system considering multiple participants
CN112883584B (zh) * 2021-03-12 2024-05-03 国网上海市电力公司 一种考虑利益交互的多能微网群冷热电多能耦合优化方法

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103730891A (zh) * 2013-11-18 2014-04-16 芜湖大学科技园发展有限公司 一种基于多智能体的微电网运行控制方法
CN104065072A (zh) * 2014-06-16 2014-09-24 四川大学 一种基于动态电价的微电网运行优化方法
CN104392279A (zh) * 2014-11-19 2015-03-04 天津大学 一种多智能体结构的微电网优化运行方法

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101510685B (zh) * 2009-03-12 2010-12-29 上海交通大学 基于生物免疫原理的分布式智能控制系统

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103730891A (zh) * 2013-11-18 2014-04-16 芜湖大学科技园发展有限公司 一种基于多智能体的微电网运行控制方法
CN104065072A (zh) * 2014-06-16 2014-09-24 四川大学 一种基于动态电价的微电网运行优化方法
CN104392279A (zh) * 2014-11-19 2015-03-04 天津大学 一种多智能体结构的微电网优化运行方法

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108075568A (zh) * 2016-11-10 2018-05-25 中国电力科学研究院 一种电力系统电压危急状态的监控系统及其构建方法
CN110111135A (zh) * 2019-04-09 2019-08-09 广东电力交易中心有限责任公司 一种发电侧成员动态报价决策方法、装置及设备
CN110111135B (zh) * 2019-04-09 2023-04-07 广东电力交易中心有限责任公司 一种发电侧成员动态报价决策方法、装置及设备
CN112215687B (zh) * 2020-11-17 2024-02-02 东北大学 一种基于云边协同的家庭微网自主竞价系统及方法
CN112215687A (zh) * 2020-11-17 2021-01-12 东北大学 一种基于云边协同的家庭微网自主竞价系统及方法
CN113377073A (zh) * 2021-06-28 2021-09-10 西南交通大学 一种基于双层多智能体系统的柔性作业车间调度优化方法
CN113570221A (zh) * 2021-07-15 2021-10-29 国网浙江省电力有限公司经济技术研究院 基于动态演化视角的电网企业拓展综合能源服务市场的辅助决策方法
CN113705067A (zh) * 2021-10-29 2021-11-26 中国电力科学研究院有限公司 一种微网优化运行策略生成方法、系统、设备及存储介质
CN113705067B (zh) * 2021-10-29 2022-02-18 中国电力科学研究院有限公司 一种微网优化运行策略生成方法、系统、设备及存储介质
CN114675535A (zh) * 2022-03-07 2022-06-28 大连理工大学 一种基于强化学习的航空发动机过渡态寻优控制方法
CN114675535B (zh) * 2022-03-07 2024-04-02 大连理工大学 一种基于强化学习的航空发动机过渡态寻优控制方法
CN114721409A (zh) * 2022-06-08 2022-07-08 山东大学 一种基于强化学习的水下航行器对接控制方法
CN116451880B (zh) * 2023-06-16 2023-09-12 华北电力大学 一种基于混合学习的分布式能源优化调度方法及装置
CN116451880A (zh) * 2023-06-16 2023-07-18 华北电力大学 一种基于混合学习的分布式能源优化调度方法及装置
CN117154954A (zh) * 2023-10-31 2023-12-01 国网浙江省电力有限公司宁波供电公司 一种基于遗传算法的配电网馈线自动化终端的配置方法
CN117154954B (zh) * 2023-10-31 2024-01-05 国网浙江省电力有限公司宁波供电公司 一种基于遗传算法的配电网馈线自动化终端的配置方法

Also Published As

Publication number Publication date
US20190108600A1 (en) 2019-04-11
CN104392279A (zh) 2015-03-04
CN104392279B (zh) 2018-02-13

Similar Documents

Publication Publication Date Title
WO2016078329A1 (zh) 一种多智能体结构的微电网优化运行方法
Tushar et al. Transforming energy networks via peer-to-peer energy trading: The potential of game-theoretic approaches
Ramachandran et al. An intelligent auction scheme for smart grid market using a hybrid immune algorithm
Faqiry et al. Double auction with hidden user information: Application to energy transaction in microgrid
Zhang et al. Energy management framework for electric vehicles in the smart grid: A three-party game
Kong et al. A multi-agent optimal bidding strategy in microgrids based on artificial immune system
WO2016078330A1 (zh) 一种基于人工免疫的微电网分布式电源竞价方法
CN112381263A (zh) 基于区块链分布式数据存储多微网日前鲁棒电能交易方法
CN109884888B (zh) 一种基于非合作博弈的多楼宇微网模型预测调控方法
Yang et al. Transactive energy system: a review of cyber‐physical infrastructure and optimal scheduling
He et al. A new cooperation framework with a fair clearing scheme for energy storage sharing
CN107220889A (zh) 一种多代理框架下微网社区的分布式资源交易方法
CN112200388A (zh) 一种基于区块链和多目标决策的p2p能源交易方法
Sun et al. Bi-level model for integrated energy service providers in joint electricity and carbon P2P market
Yang et al. Coordination and optimization of CCHP microgrid group game based on the interaction of electric and thermal energy considering conditional value at risk
Wang et al. How can the dynamic game be integrated into blockchain-based distributed energy resources multi-agent transactions for decision-making?
Wang et al. Cooperative operation of industrial/commercial/residential integrated energy system with hydrogen energy based on Nash bargaining theory
CN110556821B (zh) 考虑交互功率控制和双边竞价交易的多微电网双层优化调度方法
Wang et al. Bi-level multi-agents interactive decision-making model in regional integrated energy system
Sun et al. Nash–Cournot power market model with a high penetration of prosumers: A distributionally robust optimization approach
Li et al. Integrated energy system for low-carbon economic operation optimization: Pareto compromise programming and master-slave game
CN111126854A (zh) 一种蓄热式电采暖设备考虑使用分布式能源供电的方式
CN111080154A (zh) 基于三方博弈的市场主体在能源互联网实时市场中竞价方法
Zhang et al. A distributed coevolution algorithm for black box optimization of demand response
CN110942339B (zh) 一种虚拟电厂交易管理方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15861112

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 15861112

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 15861112

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 24.07.2018)

122 Ep: pct application non-entry in european phase

Ref document number: 15861112

Country of ref document: EP

Kind code of ref document: A1