Multi-industrial park IES optimal operation method containing multi-principal transaction and agent

Abstract

The application discloses a kind of multi-industrial park IES optimization operation methods containing multi-host transaction and agent, comprising: in each industrial park, part of energy production equipment and energy storage equipment are added in combination with the energy use characteristics of the park, while considering the flexible response resources of the electric and thermal load demand of each industrial park, forming each industrial park energy merchant to carry out energy dispatching management within each park;Build a multi-industrial park comprehensive energy system operation framework;Set the transaction operation strategy of the multi-industrial park comprehensive energy system operation framework;Define each industrial park energy merchant, energy supplier, industrial park energy operation management company, power market clearing unit and heat market clearing unit as the corresponding reinforcement learning agent;Form the multi-level transaction mode of the agent according to the set transaction operation strategy, and build a multi-level optimization operation model of each agent, output the optimal operation strategy of each agent, including transaction price and transaction energy, equipment operation output.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of integrated energy systems, and particularly relates to a multi-industrial park IES optimized operation method containing multi-agent transactions and intelligent agents. BACKGROUND

[0002] Traditional industrial parks usually obtain electric energy and thermal energy from the outside to meet the energy needs of production operation processes, daily office operations, etc. in the park, and each industrial park operates independently without energy information interaction, which results in poor energy flexibility and high operating costs of each industrial park.

[0003] With the continuous promotion of energy system digitalization reform and low-carbon transformation, intelligent industrial park energy optimization operation is attracting more and more attention. For example, photovoltaic generator sets, air source heat pumps, etc. can be established in the industrial park to meet part of the electricity, heat and other load demands in the park, which can effectively reduce the operating cost of the park. However, the energy equipment in a single industrial park may not be able to meet the energy needs of the park. At this time, energy needs to be purchased from the outside. If each industrial park purchases and sells energy independently, it will lead to high transaction prices and complex transaction processes. At this time, a corresponding energy operation manager needs to be established to manage and operate each industrial park under its jurisdiction as a representative to trade with the energy market. Therefore, how to intelligently manage the production capacity, energy use, transactions between industrial parks, and transactions between industrial parks and external energy markets is a problem that needs to be solved at present.

[0004] Based on the above technical problems, a new multi-industrial park IES optimized operation method containing multi-agent transactions and intelligent agents needs to be designed. SUMMARY

[0005] To solve the above technical problems, the technical solution of the present application is:

[0006] The present application provides a multi-industrial park IES optimized operation method containing multi-agent transactions and intelligent agents, comprising:

[0007] S1, in each industrial park, part of the production capacity equipment and energy storage equipment are added in combination with the energy characteristics of the park, and the flexible response resources of the electric and thermal load demands of each industrial park are considered to form an energy trader of each industrial park to manage and operate the energy dispatching of each park;

[0008] S2, an industrial park energy operation manager is established at the upper layer of each industrial park energy trader, which is a representative of each industrial park energy trader, and jointly constructs a multi-industrial park integrated energy system operation architecture with a power market clearing unit, an energy supplier, a heat market clearing unit and an industrial park energy trader;

[0009] S3, set the transaction operation strategy of the multi-industrial park comprehensive energy system operation architecture, comprising:

[0010] The energy trader of each industrial park outputs part of the electric energy and thermal energy to meet the electric and thermal load demand in the respective park. If the supply and demand are unbalanced, one mode or a combination of multiple modes of the preset multiple flexible modes is used to achieve supply and demand balance while achieving optimal operation economy.

[0011] The multiple flexible modes include: the first mode is a cooperative transaction between the energy traders of each industrial park; the second mode is an energy information interaction between the energy traders of each industrial park and the industrial park energy operation management company, and then the industrial park energy operation management company and the power market clearing unit and the heat market clearing unit carry out energy purchase and sale transactions; the third mode is a flexible resource response within each industrial park energy trader.

[0012] S4, define the energy traders of each industrial park, the energy suppliers, the industrial park energy operation management company, the power market clearing unit and the heat market clearing unit as corresponding reinforcement learning agents;

[0013] S5, form an agent multi-level transaction mode according to the set transaction operation strategy, and construct a multi-level optimization operation model of each agent to output the optimal operation strategy of each agent, including transaction price and transaction energy quantity, and equipment operation output.

[0014] The agent multi-level transaction mode includes: the industrial park energy operation management company agent, the energy supplier agent, and the power market clearing unit agent and the heat market clearing unit agent carry out transactions; the industrial park energy operation management company agent and each industrial park energy trader agent carry out transactions; and each industrial park energy trader agent carries out transactions.

[0015] Further, in S1, part of the energy production equipment and energy storage equipment is added in each industrial park in combination with the energy use characteristics of the park, comprising:

[0016] Obtain the electric load demand and thermal load demand change trend of each industrial park on a quarterly, monthly and daily basis. When the total electric load demand is in a high range, and the electric load demand is significantly different between the peak and valley periods, consider adding part of the electricity production equipment and energy storage equipment in the industrial park, including photovoltaic power generation equipment, wind power generation equipment and energy storage equipment. Similarly, when the total thermal load demand is in a high range, and the thermal load demand is significantly different between the peak and valley periods, consider adding part of the heat production equipment and energy storage equipment in the industrial park, including photovoltaic power generation equipment, wind power generation equipment and energy storage equipment.

[0017] Further, in the S1, considering the flexible response resources of the electricity and heat load demand of each industrial park, an energy trader of each industrial park is formed to conduct energy dispatching management within each park, including:

[0018] Considering the energy consumption period adjustment of the production and operation process of each industrial park, the transfer behavior and reduction behavior of the daily office and life operation energy consumption of the park, the demand flexible response resources and response period and response capacity of each industrial park are analyzed and reported to the energy trader of each industrial park;

[0019] The energy trader of each industrial park is an energy dispatching management platform of each industrial park, which obtains the park energy information including the operation state of the energy production and storage devices, demand response information, electricity load demand and heat load demand in each industrial park, and conducts device dispatching management and demand response execution within each industrial park.

[0020] Further, in the S2, the multi-industrial park comprehensive energy system operation architecture includes from top to bottom: the first layer is an energy supplier, the second layer is an electricity market clearing unit and a heat market clearing unit, the third layer is an industrial park energy operation management trader, and the fourth layer is an energy trader of each industrial park; the energy supplier of the first layer, the industrial park energy operation management trader of the third layer and the electricity market clearing unit and the heat market clearing unit of the second layer conduct energy purchase and sale transactions; the energy traders of each industrial park are also provided with a park electricity trading platform for realizing electricity trading between industrial parks.

[0021] Further, in the S3, the energy trader of each industrial park outputs part of the electricity and heat energy to meet the electricity and heat load demand in the park, and if the supply and demand are unbalanced, one mode or a combination of multiple modes of the preset multiple flexible modes is used to achieve supply and demand balance while achieving optimal operation economy, including:

[0022] The energy trader of each industrial park collects the historical operation parameters of the energy production and storage devices within the park, and establishes an energy supply prediction model in combination with external environmental conditions; and collects the historical energy consumption parameters within the park, and establishes a park load demand prediction model in combination with outdoor weather data, daily production and operation change conditions of the park, holiday characteristics and workday characteristics;

[0023] The energy trader of each industrial park outputs the electricity and heat supply value of each time period of each industrial park and the electricity and heat load demand value of each time period of each industrial park according to the park energy supply prediction model and the park load demand prediction model, and compares the supply and demand relationship. When the supply and demand are balanced, the power and heat load of the park is met by the production capacity equipment and the energy storage equipment in each industrial park energy trader; when the supply and demand are not balanced, a plurality of flexible modes are set in advance, according to the deviation range of the supply and demand imbalance, and the supply and demand balance and the optimal operation economy are taken as the target, one mode or a combination of multiple modes is selected, and the internal supply and demand balance of each park is realized while the optimal operation economy is achieved.

[0024] Further, when the supply and demand are not balanced, a plurality of flexible modes are set in advance, according to the deviation range of the supply and demand imbalance, and the supply and demand balance and the optimal operation economy are taken as the target, one mode or a combination of multiple modes is selected, and the internal supply and demand balance of each park is realized while the optimal operation economy is achieved.

[0025] When the supply and demand are not balanced, if the supply and demand deviation is in the low interval range, one of the third mode or the first mode is used; if the supply and demand deviation is in the middle interval range, the third mode and the first mode are combined, or the second mode is used; if the supply and demand deviation is in the high interval range, two modes or three modes are combined.

[0026] Each industrial park energy trader takes the supply and demand balance and the optimal operation economy of each park as the target, combines the mode utilization rule when the supply and demand are not balanced, and selects the optimal mode for transaction operation.

[0027] Further, the S4 defines the energy trader of each industrial park, the energy supplier, the energy operation and management trader of the industrial park, the power market clearing unit and the heat market clearing unit as the corresponding reinforcement learning agent, which comprises:

[0028] The energy trader of each industrial park is defined as the reinforcement learning agent of each industrial park: the running parameters of the production capacity equipment, the electricity and heat supply value, the electricity load and heat load demand value, the energy storage capacity of the energy storage equipment, the indoor temperature of each energy consumption area in the park, the production process running parameters, the electricity price, the heat price and the energy interaction information of other transaction subjects outside the park at t time are taken as the state space S 1,t The running output, energy transaction volume, demand response period and demand response capacity of the production capacity equipment and energy storage equipment in each industrial park at t time are taken as the action space A 1,t The state S 1,t is transferred from t time to t+1 time, and the reward function R 1,t+1 is taken as the minimum running cost of each industrial park. 1,t ​

[0029] The energy supplier is defined as a reinforcement learning agent: the state space S is defined as the energy supply and heat supply device operation parameters, electricity price, heat price, weather conditions, energy price and declared power to the market at time t 2,t The action space A is defined as the energy supply and heat supply device operation output and energy trading volume at time t 2,t The state S at time t is transferred to the state S at time t+1 2,t The state S at time t is transferred to the state S at time t+1 2,t+1 , and the reward function R is the maximum profit of the energy supplier 2,t

[0030] The industrial park energy operation manager is defined as a reinforcement learning agent: the state space S is defined as the electricity and heat supply values of all industrial parks at time t, the electricity load and heat load demand values, the energy storage capacity of energy storage devices, the time period and response capacity of demand response, and the energy price and declared power to the market 3,t The action space A is defined as the energy trading volume with the external market at time t, the energy amount allocated to each industrial park energy trader, and the demand response information 3,t The state S at time t is transferred to the state S at time t+1 3,t The state S at time t is transferred to the state S at time t+1 3,t+1 , and the reward function R is the maximum operating income of the industrial park energy operation manager 3,t

[0031] The power market clearing unit and the heat market clearing unit are defined as reinforcement learning agents: the state space S is defined as the energy price and power declared by the energy supplier and the energy price and power declared by the industrial park energy operation manager at time t 4,t The action space A is defined as the clearing price, clearing heat price and clearing quantity at time t 4,t The state S at time t is transferred to the state S at time t+1 4,t The state S at time t is transferred to the state S at time t+1 4,t+1 , and the reward function R is the maximum social welfare 4,t .

[0032] Further, in the S5, the industrial park energy operation manager agent, the energy supplier agent, and the power market clearing unit agent, the heat market clearing unit agent trade, including:

[0033] At the beginning of the initial bidding, the industrial park energy operation manager agent and the energy supplier agent respectively declare their own energy price and power to the power market clearing unit agent and the heat market clearing unit agent, and then the power market clearing unit agent and the heat market clearing unit agent process the declared data according to the bidding clearing mechanism, and output the clearing data information; ​​

[0034] The power market dispatch unit agent and the heat market dispatch unit agent are leaders, and the industrial park energy operation management agent and the energy supplier agent are followers, and the dispatch data information is analyzed according to the game theory market benefit balance, whether it can be directly dispatched, if the benefits are not balanced, the benefit imbalance signal is sent to the industrial park energy operation management agent and the energy supplier agent, and the final dispatch data information is output until the benefit balance is output according to the dispatch data information;

[0035] The industrial park energy operation management agent and each industrial park energy agent trade, including:

[0036] The industrial park energy operation management agent is a leader, and each industrial park energy agent is a follower, the leader formulates a pricing strategy according to the transaction information reported by each follower, and publishes it to each follower, each follower adjusts its actual energy demand according to the price signal, re-reports the transaction information, and builds a master-slave game model and solves it to output the transaction information between the two.

[0037] The industrial park energy operation management agent and each industrial park energy agent trade, including: defining the transaction between each industrial park energy agent as a cooperative game transaction, taking the minimum difference between the cost of each industrial park energy agent before and after participating in Nash bargaining as the target, and taking the transaction price between each industrial park energy agent as the decision variable, constructing a cooperative game Nash bargaining model, and solving and outputting the transaction information between each industrial park energy agent.

[0038] Further, in S5, a multi-layer optimization operation model of each agent is constructed, including:

[0039] According to the transaction process of the industrial park energy operation management agent, the energy supplier agent, the power market dispatch unit agent and the heat market dispatch unit agent, the transaction and optimization operation objects are determined, and the optimization objective function of each object is formed in combination with the reward function of the corresponding agent, and then a first layer master-slave game optimization operation model is constructed, which is expressed as:

[0040]

[0041] (IEOM∪ES∪EMC∪HMC) is a participant in the game, including the industrial park energy operation management agent, the energy supplier agent, the power market dispatch unit agent and the heat market dispatch unit agent; P and P respectively represent the transaction price and heat price strategy output by the power market dispatch unit and the heat market dispatch unit to the industrial park energy operation management agent and the energy supplier agent; P t e P th respectively are the transaction electricity and heat quantity strategies of the industrial park energy operation management company, the energy supplier and the power market and heat market clearing units; is the operation output strategy of the i-th device in the energy supplier; I IEOM ES EMC HMC respectively are the optimization objective functions of the industrial park energy operation management company, the energy supplier and the power market and heat market clearing units in the transaction;

[0042] According to the transaction process between the industrial park energy operation management company agent and each industrial park energy company agent, the transaction and optimization operation objects are determined, and the optimization objective functions of each object are formed in combination with the reward functions of the corresponding agents, and then a second-layer master-slave game optimization operation model is constructed, which is expressed as:

[0043]

[0044] (IEOM∪IE1∪IE2∪…∪IE n ) are the participants of the game, including the industrial park energy operation management company and n industrial park energy companies; respectively are the transaction electricity and heat price strategies of the industrial park energy operation management company to the industrial park energy companies; respectively are the transaction electricity and heat quantity strategies between each industrial park energy company and the industrial park energy operation management company; is the operation output strategy of the j-th device in the industrial park energy company; P t DR is the flexible resource demand response strategy in the industrial park energy company; I' IEOM respectively are the optimization objective functions of the industrial park energy operation management company and each industrial park energy company in the transaction;

[0045] According to the cooperative game Nash bargaining model between each industrial park energy company agent, the transaction electricity price is determined, and the optimization objective is formed in combination with the reward functions of the corresponding agents, and then a third-layer cooperative game optimization operation model is constructed, which is expressed as:

[0046]

[0047] C e,k,t is the power generation cost of the k-th industrial park energy company at the t time period; C h,k,t is the heat generation cost of the k-th industrial park energy company at the t time period; C es,k,t is the energy storage cost of the k-th industrial park energy company at the t time period; C m,k,t ​​​​The energy transaction cost of the kth industrial park energy trader at the t period.

[0048] Further, the solving method of the first layer master-slave game optimization operation model, the second layer master-slave game optimization operation model and the third layer cooperative game optimization operation model adopts a deep reinforcement learning algorithm.

[0049] The beneficial effects of the present application are:

[0050] (1) The present application adds part of energy production equipment and energy storage equipment in each industrial park in combination with the energy use characteristics of the park, simultaneously considers the flexible response resources of the electricity and heat load demand of each industrial park, forms the energy dispatching management of each industrial park energy trader, establishes an industrial park energy operation management trader at the upper layer of each industrial park energy trader, as the representative of each industrial park energy trader, and jointly constructs a multi-industrial park comprehensive energy system operation framework with the power market clearing unit, the energy supplier, the heat market clearing unit and the industrial park energy trader; through the addition of energy production equipment and energy storage equipment, the equipment can be flexibly adjusted according to the energy use characteristics of the park, energy waste is reduced, and the demand flexible response resources are considered, the industrial park can be flexibly adjusted according to the energy price and the production operation process; in addition, the establishment of the industrial park energy operation management trader can realize the overall scheduling and management of each industrial park energy trader, as the representative to trade with the power market clearing unit and the heat market clearing unit, which is convenient for energy purchase and sale transactions with the market clearing unit when the supply and demand are imbalanced in the industrial park energy trader;

[0051] (2) When each industrial park energy trader is an energy producer, part of the output electric energy and heat energy meets the electricity and heat load demand in each park, if the supply and demand are imbalanced, one mode or a combination of multiple modes of the preset multiple flexible modes is used to achieve supply and demand balance and optimal operation economy; the supply and demand balance in the industrial park and the minimum park operation cost can be realized through multiple flexible modes under the condition of imbalance in the industrial park;

[0052] (3) the present application defines each industrial park energy trader, energy supplier, industrial park energy operation manager, power market clearing unit, heat market clearing unit as the corresponding reinforcement learning agent; according to the set transaction operation strategy, the multi-level transaction mode of the agent is formed, and the multi-level optimization operation model of each agent is constructed, and the optimal operation strategy of each agent is output, including transaction price and transaction energy, equipment operation output; it can effectively utilize the reinforcement learning agent to quickly adapt to the change of environment and the uncertainty of system, such as price fluctuation, supply and demand relationship change, equipment operation state, etc., each agent automatically learns the optimal decision strategy, ensures the competitive advantage of each agent in the complex and changeable environment information, improves the accuracy and efficiency of decision-making, reduces the uncertainty of human intervention; in addition, the multi-level transaction mode makes different agents can more effectively communicate and cooperate, realize the optimal scheduling management of resources and energy, and use the master-slave game, cooperative game optimization operation model in the multi-level optimization model, through layer-by-layer optimization decision, realize the economic optimization of each agent.

[0053] Other features and advantages will be set forth in the following description of the application, and in part will become apparent to those skilled in the art upon examination of the following or can be learned by practice of the application. The objects and other advantages of the application will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.

[0054] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the following preferred embodiments are specifically described, and the accompanying drawings are described in detail as follows. BRIEF DESCRIPTION OF DRAWINGS

[0055] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings needed to be used in the specific embodiments or the prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0056] Figure 1 A flow chart of a multi-industrial park IES optimization operation method containing multiple agents and agents according to the present application;

[0057] Figure 2 A multi-industrial park IES structure diagram according to the present application;

[0058] Figure 3 A first layer master-slave game principle diagram according to the present application;

[0059] Figure 4 A second layer master-slave game and third layer cooperative game principle diagram according to the present application. DETAILED DESCRIPTION

[0060] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions of the present application will be described clearly and completely below with reference to the drawings. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0061] Embodiment 1

[0062] As shown in Figure 1 , Figure 2 , the present embodiment 1 provides a multi-industrial park IES optimization operation method containing multi-agent transaction and agent, comprising:

[0063] S1, in each industrial park, part of energy production equipment and energy storage equipment are added in combination with the energy use characteristics of the park, and the flexible response resources of the electricity and heat load demand of each industrial park are considered to form an energy trader of each industrial park to carry out energy dispatching management within each park;

[0064] S2, an industrial park energy operation management trader is established at the upper layer of each industrial park energy trader, which serves as a representative of each industrial park energy trader, and together with the power market clearing unit, energy supplier, heat market clearing unit and industrial park energy trader, forms a multi-industrial park comprehensive energy system operation architecture;

[0065] S3, a transaction operation strategy of the multi-industrial park comprehensive energy system operation architecture is set, comprising:

[0066] When each industrial park energy trader is an energy producer, it outputs part of the electric energy and heat energy to meet the electricity and heat load demand in the park, and if the supply and demand are unbalanced, one mode or a combination of multiple modes of the preset multiple flexible modes is used to achieve supply and demand balance while achieving optimal operation economy;

[0067] Among them, the multiple flexible modes include: the first mode is a cooperative transaction between each industrial park energy trader; the second mode is an energy information interaction between each industrial park energy trader and the industrial park energy operation management trader, and then a purchase and sale energy transaction is carried out between the industrial park energy operation management trader and the power market clearing unit and the heat market clearing unit; the third mode is a flexible resource response within each industrial park energy trader;

[0068] S4, each industrial park energy trader, energy supplier, industrial park energy operation management trader, power market clearing unit and heat market clearing unit are defined as corresponding reinforcement learning agents;

[0069] S5, forming the multi-level trading mode of the agent according to the set transaction operation strategy, and constructing a multi-level optimization operation model of each agent to output the optimal operation strategy of each agent, including the transaction price and the transaction energy amount, and the device operation output;

[0070] The multi-level trading mode of the agent includes: the industrial park energy operation management agent, the energy supplier agent, and the power market clearing unit agent and the heat market clearing unit agent trade with each other; the industrial park energy operation management agent trades with each industrial park energy agent; and each industrial park energy agent trades with each other.

[0071] In the embodiment, in S1, part of the energy production equipment and energy storage equipment are added in each industrial park according to the energy consumption characteristics of the park, including:

[0072] The change trend of the electric load demand and the heat load demand of each industrial park is obtained. When the total electric load demand is in a high range, and the electric load demand is obviously different between the peak and the valley, part of the power generation equipment and energy storage equipment, including photovoltaic power generation equipment, wind power generation equipment, and energy storage equipment, are added in the industrial park. Similarly, when the total heat load demand is in a high range, and the heat load demand is obviously different between the peak and the valley, part of the heat production equipment and energy storage equipment, including photovoltaic power generation equipment, wind power generation equipment, and energy storage equipment, are added in the industrial park.

[0073] In actual application, because of the different park scales, production processes, and daily office operation modes of each industrial park, the change trend of the electric load demand and the heat load demand of each industrial park is also different. Moreover, the electric load demand of some industrial parks is high, and the heat load demand is low. Usually, only heat load demand exists in winter. Such industrial parks are defined as power consumption type industrial parks. It is necessary to establish power generation equipment and energy storage equipment in such parks to save the cost of purchasing electricity. In addition, because of the operation requirements of the production process, the heat load demand of some industrial parks is high, and the electric load demand mainly concentrates on daily personnel office. Therefore, it is necessary to establish heat production equipment and heat storage equipment near the production workshop. According to the park scale and economic conditions, power generation equipment and energy storage equipment can also be established. It is also feasible to convert electric energy into heat energy. Therefore, each industrial park needs to add energy production equipment and energy storage equipment according to the actual operation situation and economic conditions of the park.

[0074] In the embodiment, in S1, the flexible response resources of the electric load demand and the heat load demand of each industrial park are considered to form the energy dispatching management of each park energy agent, including:

[0075] Consider the energy consumption period adjustment of the production operation process of each industrial park, the transfer behavior and reduction behavior of the daily office and life operation energy consumption of the park, analyze the demand flexible response resources and response period, response capacity of each industrial park, and report to the energy trader of each industrial park;

[0076] The energy trader of each industrial park is the energy dispatching and management platform of each industrial park, obtains the park energy information including the running state of the production and energy storage devices in each industrial park, demand response information, and electric and thermal load demand, and performs device dispatching and management and demand response execution in each industrial park.

[0077] It should be noted that some industrial parks have production off-season and production peak season, and the energy consumption is small in the production off-season. The appropriate energy price period can be selected or the production can be carried out when the production capacity is relatively large, so that the energy consumption behavior transfer and reduction are realized. Or the production operation process of some industrial parks is automatically operated, and the running time of the related production process can be set according to the production capacity in the park, so that the energy consumption period is flexible and the running economy is optimal.

[0078] In the embodiment, in S2, the multi-industrial park comprehensive energy system operation architecture includes from top to bottom: the first layer is an energy supplier, the second layer is an electric power market clearing unit and a heat market clearing unit, the third layer is an industrial park energy operation management trader, and the fourth layer is each industrial park energy trader; the energy supplier of the first layer, the industrial park energy operation management trader of the third layer, and the electric power market clearing unit and the heat market clearing unit of the second layer carry out energy purchase and sale transactions; the park electric power transaction platform is further arranged between the energy traders of each industrial park, and is used to realize electric power transaction between the industrial parks.

[0079] In the embodiment, in S3, when the energy trader of each industrial park is an energy producer, part of the electric and thermal energy is output to meet the electric and thermal load demand in the park. If the supply and demand are unbalanced, one mode or a combination of multiple modes of the preset multiple flexible modes is used to achieve supply and demand balance and optimal running economy, including:

[0080] The energy trader of each industrial park collects the historical running parameters of the production and energy storage devices in the park, and establishes a park energy supply prediction model in combination with external environmental conditions; and collects the energy consumption historical parameters in the park, and establishes a park load demand prediction model in combination with outdoor weather data, daily production and operation change conditions of the park, holiday characteristics, and working day characteristics;

[0081] The energy trader of each industrial park outputs the electricity and heat supply value of each time period of each industrial park and the electricity and heat load demand value of each time period of each industrial park according to the park energy supply prediction model and the park load demand prediction model, and compares the supply and demand relationship. When the supply and demand are balanced, the power and heat load of the park is met by the production capacity equipment and energy storage equipment in the energy trader of each industrial park. When the supply and demand are not balanced, a plurality of flexible modes are set in advance, and according to the deviation range of the supply and demand imbalance, and taking the supply and demand balance and the optimal operation economy as the target, one mode or a combination of multiple modes is selected for use, so as to achieve the supply and demand balance of each park while achieving the optimal operation economy.

[0082] In the embodiment, when the supply and demand are not balanced, a plurality of flexible modes are set in advance, and according to the deviation range of the supply and demand imbalance, and taking the supply and demand balance and the optimal operation economy as the target, one mode or a combination of multiple modes is selected for use, which includes:

[0083] When the supply and demand are not balanced, if the supply and demand deviation is in the low interval range, one of the third mode or the first mode is used; if the supply and demand deviation is in the middle interval range, the third mode and the first mode are combined, or the second mode is used; if the supply and demand deviation is in the high interval range, two modes or three modes are combined.

[0084] The energy trader of each industrial park takes the supply and demand balance and the optimal operation economy of each park as the target, and selects the optimal mode for transaction operation according to the mode utilization rule when the supply and demand are not balanced.

[0085] In actual application, when the energy supply in the energy trader of an industrial park is greater than the demand, the supply and demand deviation is in the low interval, the excess energy can be stored by using the energy storage device, or the excess energy can be used to increase the production process operation output, or the excess energy can be sold to other industrial park energy traders. When the supply is less than the demand, the supply and demand deviation is in the low interval, the missing energy can be compensated by using the energy storage resource, or the production process operation output can be reduced, or the missing energy can be purchased from other industrial park energy traders. In addition, when the supply and demand deviation is in the middle interval, a combination of multiple modes can be used to compensate for the deviation. For example, through the combination of the first mode and the third mode, or the mode combination brings more complexity to the production operation process adjustment, the energy market clearing unit and the heat market clearing unit can be purchased or sold by the industrial park energy operation management trader. When the supply and demand deviation is in the high interval, multiple modes can be flexibly combined and applied to achieve the supply and demand balance of the park, and also ensure the operation economy of the park.

[0086] In the embodiment, the S4 defines the energy trader of each industrial park, the energy supplier, the energy operation and management trader of the industrial park, the power market clearing unit and the heat market clearing unit as the corresponding reinforcement learning agent, comprising:

[0087] The energy trader of each industrial park is defined as the reinforcement learning agent of each industrial park: the running parameters of the production capacity equipment, the power and heat supply values, the power load and heat load demand values, the energy storage capacity of the energy storage equipment, the indoor temperature of each energy consumption area in the park, the production process running parameters, the electricity price, the heat price and the energy interaction information of other transaction subjects outside the park at time t are taken as the state space S 1,t The running output of the production capacity equipment and the energy storage equipment in each industrial park at time t, the energy transaction volume, the demand response period and the demand response capacity are taken as the action space A 1,t The state S 1,t at time t is transferred to the state S 1,t+1 at time t+1, and the minimum running cost of each industrial park is taken as the reward function R 1,t ;

[0088] The energy supplier is defined as the reinforcement learning agent: the running parameters of the power and heat supply equipment at time t, the electricity price, the heat price, the weather condition, the energy price and the declared power reported to the market are taken as the state space S 2,t The running output of the power and heat supply equipment at time t, the energy transaction volume are taken as the action space A 2,t The state S 2,t at time t is transferred to the state S 2,t+1 at time t+1, and the maximum profit of the energy supplier is taken as the reward function R 2,t ;

[0089] The energy operation and management trader of the industrial park is defined as the reinforcement learning agent: the power and heat supply values of all industrial parks at time t, the power load and heat load demand values, the energy storage capacity of the energy storage equipment, the period and response capacity participating in the demand response, the energy price and the declared power reported to the market are taken as the state space S 3,t The energy transaction volume with the external market at time t, the energy volume allocated to the energy trader of each industrial park, the demand response information are taken as the action space A 3,t The state S 3,t at time t is transferred to the state S 3,t+1 at time t+1, and the maximum running income of the energy operation and management trader of the industrial park is taken as the reward function R 3,t ;

[0090] The power market clearing unit and the heat market clearing unit are defined as reinforcement learning agents: the energy price and power declared by the energy supplier at time t and the energy price and power declared by the industrial park energy operation management company are taken as the state space S 4,t The clearing price, the clearing heat price and the clearing quantity at time t are taken as the action space A 4,t The state S 4,t transferred from time t to time t+1 is taken as the state S 4,t+1 , and the maximum social welfare is taken as the reward function R 4,t .

[0091] It should be noted that the reinforcement learning task is generally described by a Markov decision process, which corresponds to a four-tuple of state space S, action space A, state transition matrix P and reward function R. The agent Agent is in the environment, and the environment perceived by the agent is called the state; the actions that the agent Agent can take constitute the action space; when a certain action acts on the current state, the environment will be transferred from the current state to another state according to the state transition matrix, and the environment will feed back the corresponding reward to the agent Agent according to the reward function. The agent Agent obtains the corresponding action strategy by continuously interacting with the environment, and the advantages and disadvantages of the action strategy depend on the cumulative reward obtained after long-term execution of the strategy.

[0092] The energy operation management of each industrial park energy company agent can be regarded as an environment changing optimization problem, and the goal of the agent is to determine the action decision to be executed according to the own goal through the interaction information with the environment, and then minimize the operation cost of the industrial park through the optimization of the internal production capacity equipment, energy storage equipment and flexible resources. The industrial park energy company needs to interact with other industrial park energy companies, and also needs to interact with the power market clearing unit and the heat market clearing unit, and there are external energy suppliers and power market clearing units and heat market clearing units, so the corresponding agents of the energy suppliers, industrial park energy operation management companies, power market clearing units and heat market clearing units also need to be established. The environment, state and reward function of each agent are different, but the information interaction between the agents is needed, so the related information of other agents in the state space is also involved when the agents optimize the operation, and a multi-agent multi-industrial park comprehensive energy system operation architecture is formed.

[0093] As shown in Figure 3 , Figure 4 In the S5, the industrial park energy operation management company agent, the energy supplier agent, and the power market clearing unit agent and the heat market clearing unit agent trade, including:

[0094] At the beginning of the initial offer, the industrial park energy operation management agent and the energy supplier agent respectively declare the energy price and power to the power market clearing unit agent and the heat market clearing unit agent, and then the power market clearing unit agent and the heat market clearing unit agent process the declaration data according to the clearing mechanism to output the clearing data information;

[0095] The power market clearing unit agent and the heat market clearing unit agent are taken as leaders, and the industrial park energy operation management agent and the energy supplier agent are taken as followers. The clearing data information is analyzed according to the game theory market benefit balance. Whether it can be directly cleared, if the benefits are not balanced, the industrial park energy operation management agent and the energy supplier agent are sent a benefit imbalance signal. According to the clearing data information, the re-offer is made until the final clearing data information is outputted.

[0096] The industrial park energy operation management agent and the energy supplier agent are taken as leaders, and the industrial park energy operation management agent and the energy supplier agent are taken as followers. The clearing data information is analyzed according to the game theory market benefit balance. Whether it can be directly cleared, if the benefits are not balanced, the industrial park energy operation management agent and the energy supplier agent are sent a benefit imbalance signal. According to the clearing data information, the re-offer is made until the final clearing data information is outputted.

[0097] The industrial park energy operation management agent is taken as a leader, and the industrial park energy supplier agent is taken as a follower. The leader makes a pricing strategy according to the transaction information declared by each follower and publishes it to each follower. Each follower adjusts its actual energy demand according to the price signal, re-declares the transaction information, and builds a master-slave game model and solves it to output the transaction information between the two.

[0098] The industrial park energy operation management agent and the energy supplier agent are taken as leaders, and the industrial park energy operation management agent and the energy supplier agent are taken as followers. The clearing data information is analyzed according to the game theory market benefit balance. Whether it can be directly cleared, if the benefits are not balanced, the industrial park energy operation management agent and the energy supplier agent are sent a benefit imbalance signal. According to the clearing data information, the re-offer is made until the final clearing data information is outputted.

[0099] It should be noted that the Nash bargaining theory belongs to the range of cooperative game, which can help decision makers achieve fair Pareto optimal benefit distribution. In the case of participants with interactive characteristics and conflicting interests, Nash bargaining can take into account the overall and individual interests.

[0100] The cooperative game Nash bargaining model is expressed as:

[0101]

[0102] The cost of the industrial park energy supplier k not participating in the bargaining when the industrial park energy suppliers are in cooperation alliance; The cost of the industrial park energy trader k participating in the price negotiation when the industrial park energy traders form a cooperative alliance; The optimal cost when the industrial park energy traders form a cooperative alliance; The electricity selling revenue of the industrial park energy trader k to other industrial park energy traders when the industrial park energy traders form a cooperative alliance; The transaction electricity price between the industrial park energy traders when the industrial park energy traders form a cooperative alliance; The optimal transaction electricity quantity between the industrial park energy traders when the industrial park energy traders form a cooperative alliance.

[0103] In the embodiment, the S5 includes constructing a multi-layer optimization operation model of each agent, including:

[0104] According to the transaction process of the industrial park energy operation management agent, the energy supply agent, the power market clearing unit agent and the heat market clearing unit agent, the transaction and optimization operation objects are determined, and the optimization objective functions of each object are formed in combination with the reward functions of the corresponding agents, and then a first-layer master-slave game optimization operation model is constructed, which is expressed as:

[0105]

[0106] (IEOM∪ES∪EMC∪HMC) are the participants of the game, including the industrial park energy operation management agent, the energy supply agent, the power market clearing unit and the heat market clearing unit; respectively represent the transaction electricity price and heat price strategies output by the power market clearing unit and the heat market clearing unit to the industrial park energy operation management agent and the energy supply agent; P t e ,P t h respectively are the transaction electricity quantity and heat quantity strategies of the industrial park energy operation management agent, the energy supply agent, the power market clearing unit and the heat market clearing unit; is the operation output strategy of the i-th device in the energy supply agent; I IEOM , I ES , I EMC , I HMC respectively are the optimization objective functions of the industrial park energy operation management agent, the energy supply agent, the power market clearing unit and the heat market clearing unit when they trade with each other;

[0107] According to the transaction process of the industrial park energy operation management agent and each industrial park energy agent, the transaction and optimization operation objects are determined, and the optimization objective functions of each object are formed in combination with the reward functions of the corresponding agents, and then a second-layer master-slave game optimization operation model is constructed, which is expressed as:

[0108]

[0109] (IEOM∪IE1∪IE2∪…∪IE n ) are participants of the game, including the industrial park energy operation manager and n industrial park energy traders; respectively are the transaction electricity price and transaction heat price strategies output by the industrial park energy operation manager to the industrial park energy traders; respectively are the transaction electricity and transaction heat strategies between each industrial park energy trader and the industrial park energy operation manager; is the operation output strategy of the jth device in the industrial park energy trader; P t DR is the flexible resource demand response strategy in the industrial park energy trader; I I EOM , respectively are the optimization objective functions of the industrial park energy operation manager and each industrial park energy trader when trading;

[0110] According to the cooperative game Nash bargaining model between the industrial park energy trader agents, the transaction electricity price is determined, and the optimization objective is formed in combination with the reward function of the corresponding agent, and then a third layer cooperative game optimization operation model is constructed, which is expressed as:

[0111]

[0112] C e,k,t is the electricity generation cost of the kth industrial park energy trader at the t period; C h,k,t is the heat generation cost of the kth industrial park energy trader at the t period; C es,k,t is the energy storage cost of the kth industrial park energy trader at the t period; C m,k,t is the electricity transaction cost of the kth industrial park energy trader at the t period.

[0113] In this embodiment, the solving method of the first layer master-slave game optimization operation model, the second layer master-slave game optimization operation model and the third layer cooperative game optimization operation model adopts a deep reinforcement learning algorithm.

[0114] ​In several embodiments provided in the present application, it should be understood that the disclosed system and method can also be implemented by other manners. The system embodiments described above are only illustrative, for example, the flowcharts and block diagrams in the drawings show the possible implementation architecture, function and operation of the system, method and computer program product according to the embodiments of the present application. In this regard, each block in the flowchart or block diagram can represent a module, a program segment or a part of code, which includes one or more executable instructions for implementing the specified logic function. It should also be noted that in some alternative implementations, the functions noted in the blocks can occur in different order from that shown in the drawings. For example, two consecutive blocks can actually be executed substantially in parallel, and sometimes they can be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0115] In addition, each functional module in the embodiments of the present application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part. When the functions are realized in the form of software functional modules and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the parts that contribute to the prior art or parts of the technical solutions can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the embodiments of the method of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk or an optical disk, and various media that can store program codes.

[0116] Based on the above ideal embodiments according to the present application, through the above description, relevant personnel can make various changes and modifications without deviating from the scope of the technical idea of the present application. The technical scope of the present application is not limited to the contents in the specification, and must be determined by the scope of the claims.

Claims

1. A method for optimizing the operation of an industrial park IES (Environmentally Integrated Systems) involving multiple entities and intelligent agents, characterized in that: include: S1. In each industrial park, some production capacity equipment and energy storage equipment are added according to the energy consumption characteristics of the park. At the same time, the power and heat load demand of each industrial park is considered to flexibly respond to resources, so as to form an energy supplier in each industrial park to carry out energy dispatch management within each park. S2. Establish an industrial park energy operation and management provider above the energy providers in each industrial park, acting as the representative of the energy providers in each industrial park, and jointly construct a multi-industrial park integrated energy system operation architecture with the electricity market clearing unit, energy suppliers, heat market clearing unit, and industrial park energy providers. S3. A trading operation strategy for setting up an integrated energy system operation architecture across multiple industrial parks, including: When energy providers in each industrial park act as energy producers, they output a portion of electrical and thermal energy to meet the electricity and heat load demands within their respective parks. If there is an imbalance between supply and demand, they utilize one or a combination of preset flexible modes to achieve optimal operational economy while maintaining supply and demand balance. The various flexible modes include: the first mode is cooperative transactions between energy providers in various industrial parks; the second mode is energy information exchange between energy providers in various industrial parks and energy operation and management providers in industrial parks, and then energy purchase and sale transactions between energy operation and management providers in industrial parks and electricity market clearing units and heat market clearing units; the third mode is flexible resource response within each energy provider in various industrial parks. S4. Define the energy providers, energy suppliers, energy operators, electricity market clearing units, and heat market clearing units of each industrial park as corresponding reinforcement learning agents. S5. Based on the set trading operation strategy, form a multi-level trading mode for intelligent agents, construct a multi-level optimized operation model for each intelligent agent, and output the optimal operation strategy for each intelligent agent, including trading price, trading energy quantity, and equipment operation output. The multi-level transaction method of the intelligent agents includes: transactions between the intelligent agents of industrial park energy operation and management, the intelligent agents of energy suppliers, and the intelligent agents of electricity market clearing units and heat market clearing units; transactions between the intelligent agents of industrial park energy operation and management and the intelligent agents of energy suppliers in various industrial parks; and transactions between the intelligent agents of energy suppliers in various industrial parks. In S5, a multi-layered optimization operation model for each agent is constructed, including: Based on the transaction process between the intelligent agents of the industrial park energy operation management provider, the intelligent agent of the energy supplier, and the intelligent agents of the electricity market clearing unit and the intelligent agent of the heat market clearing unit, the transaction and optimization operation objects are determined, and the optimization objective function of each object is formed by combining the reward function of the corresponding intelligent agent. Then, the first-level master-slave game optimization operation model is constructed, which is expressed as: ； The participants in the game include industrial park energy operators, energy suppliers, electricity market clearing units, and heat market clearing units; , These represent the electricity and heat pricing strategies that the electricity market clearing unit and the heat market clearing unit provide to the energy operators and suppliers of the industrial park, respectively. , These are the trading volume and heat strategies for energy operators and managers, energy suppliers, and electricity market clearing units and heat market clearing units in industrial parks, respectively. The power output strategy for the i-th device within the energy supplier; , , , These are the respective optimization objective functions for energy operators and managers in industrial parks, energy suppliers, and electricity and heat market clearing units when trading; Based on the transaction process between the intelligent agents of industrial park energy operation and management and the intelligent agents of various industrial park energy providers, the transaction and optimization operation objects are determined, and the optimization objective function of each object is formed by combining the reward function of the corresponding intelligent agent. Then, a second-layer master-slave game optimization operation model is constructed, which is expressed as: ； The participants in the game include the energy operation and management company of the industrial park and n energy companies in each industrial park; , These are the trading electricity price and trading heat price strategies that the energy operation and management company of the industrial park provides to the energy companies in the industrial park; , These are the trading strategies for electricity and heat between energy providers and energy operators in various industrial parks; The power output strategy for the j-th equipment in the industrial park's energy sector; A flexible resource demand response strategy for energy providers within industrial parks; , These are the respective optimization objective functions for the energy operation and management company of the industrial park and the energy providers of each industrial park when they trade; Based on the Nash bargaining model of cooperative game theory among energy agents in various industrial parks, the transaction price of electricity is determined, and the optimization objective is formed by combining the reward function of the corresponding agents. Then, a third-layer cooperative game optimization operation model is constructed, expressed as: ； Let be the electricity generation cost of the energy provider in the k-th industrial park during time period t; The heat production cost of the energy provider in the kth industrial park during time period t; Let be the energy storage cost of the energy provider in the k-th industrial park during time period t; Let be the electricity transaction cost of the energy provider in the k-th industrial park during time period t.

2. The multi-industrial park IES optimized operation method according to claim 1, characterized in that, In S1, some production capacity equipment and energy storage equipment are added to each industrial park according to the park's energy consumption characteristics, including: Obtain the quarterly, monthly, and daily trends of electricity and heat load demand in each industrial park. When the total electricity load demand is in a high range and there is a significant difference between the peak and off-peak periods, consider adding some power generation and energy storage equipment within the industrial park, including photovoltaic power generation equipment, wind power generation equipment, and energy storage equipment. Similarly, when the total heat load demand is in a high range and there is a significant difference between the peak and off-peak periods, consider adding some heat generation and energy storage equipment within the industrial park, including solar thermal power generation equipment, electric heat pumps, and energy storage and thermal storage equipment.

3. The multi-industrial park IES optimized operation method according to claim 1, characterized in that, In S1, considering the flexible response resources for the electricity and heat load demands of each industrial park, energy providers in each industrial park conduct energy dispatch management within each park, including: Considering the energy consumption time-limited adjustability of production and operation processes in each industrial park, as well as the transfer and reduction behaviors of daily office and living operations in the parks, analyze the demand-flexible response resources, response time periods, and response capacity of each industrial park, and report to the energy suppliers of each industrial park. Each industrial park energy provider acts as the energy dispatch and management platform for its respective industrial park. It acquires energy information for the park, including the operating status of production equipment and energy storage equipment, demand response information, electricity load demand, and heat load demand. It also conducts equipment dispatch management and demand response execution within each industrial park.

4. The multi-industrial park IES optimized operation method according to claim 1, characterized in that, In S2, the multi-industrial park integrated energy system operation architecture includes, from top to bottom: the first layer is the energy supplier, the second layer is the electricity market clearing unit and the heat market clearing unit, the third layer is the industrial park energy operation and management unit, and the fourth layer is the energy supplier of each industrial park; the energy supplier of the first layer, the industrial park energy operation and management unit of the third layer and the electricity market clearing unit and the heat market clearing unit of the second layer conduct energy purchase and sale transactions; the energy suppliers of each industrial park are also connected by a park electricity trading platform to realize electricity transactions between the various industrial parks.

5. The multi-industrial park IES optimized operation method according to claim 1, characterized in that, In S3, each industrial park energy provider, acting as an energy producer, outputs a portion of its electrical and thermal energy to meet the electricity and heat load demands within its respective park. If supply and demand are unbalanced, one or more preset flexible modes are used to achieve optimal operational economy while maintaining supply and demand balance. These modes include: Energy providers in each industrial park collect historical operating parameters of their production capacity and energy storage equipment, and combine them with external environmental conditions to establish an energy supply forecasting model for the park; they also collect historical energy consumption parameters of their park, and combine them with outdoor weather data, changes in daily production and operation conditions, and holiday and weekday characteristics to establish a load demand forecasting model for the park. Each industrial park energy provider outputs the electricity and heat supply values ​​and the electricity and heat load demand values ​​for each industrial park at each time period based on the park's energy supply forecasting model and park load demand forecasting model. After comparing the supply and demand relationship, when the supply and demand are balanced, the power generation equipment and energy storage equipment in each industrial park energy provider directly meet the park's electricity and heat load. When the supply and demand are unbalanced, multiple flexible modes are preset. Based on the deviation range of the supply and demand imbalance, and with the goal of achieving supply and demand balance and optimal operating economy, one or more of these modes are selected to achieve optimal operating economy while achieving supply and demand balance within each park.

6. The multi-industrial park IES optimized operation method according to claim 5, characterized in that, When supply and demand are unbalanced, multiple flexible modes are pre-set. Based on the range of the supply-demand imbalance and with the goal of achieving supply-demand balance and optimal operational economy, one or more of these modes are selected for use, including: When supply and demand are unbalanced, if the supply and demand deviation is in the low range, then use the third mode or one of the first modes; if the supply and demand deviation is in the middle range, then use a combination of the third mode and the first mode, or use the second mode; if the supply and demand deviation is in the high range, then use a combination of two of the first mode, the second mode and the third mode, or a combination of all three modes. Each industrial park energy provider aims to achieve optimal supply and demand balance and operational economics within its own park. It combines the rules for handling supply and demand imbalances to select the optimal trading model.

7. The multi-industrial park IES optimized operation method according to claim 1, characterized in that, S4 defines the energy providers, energy suppliers, energy operators, electricity market clearing units, and heat market clearing units of each industrial park as corresponding reinforcement learning agents, including: Each industrial park energy provider is defined as a reinforcement learning agent for that industrial park: the operating parameters of production equipment, electricity and heat supply, electricity and heat demand, energy storage capacity of energy storage devices, indoor temperature of each energy-consuming area within the park, production process operating parameters, electricity price, heat price, and energy interaction information of other external trading entities at time t are used as the state space. The operating output, energy trading volume, demand response period, and demand response capacity of the production capacity and energy storage equipment within each industrial park at time t are used as the action space. The state at time t Transition to the state at time t+1 And the reward function is the minimum operating cost of each industrial park. ; The energy supplier is defined as a reinforcement learning agent: the operating parameters of the electrical and thermal energy supply equipment at time t, the electricity price, the thermal price, the weather conditions, the energy price declared to the market, and the declared power are used as the state space. The operating output of electrical and thermal energy supply equipment and the amount of energy traded at time t are used as the action space. The state at time t Transition to the state at time t+1 And the reward function is based on maximizing the profits of energy suppliers. ; The energy operation and management provider of the industrial park is defined as a reinforcement learning intelligent agent: the electricity and heat supply values, electricity and heat load demand values, energy storage capacity of energy storage devices, time periods and response capacities for participating in demand response, energy prices submitted to the market, and submitted power at time t are used as the state space of all industrial parks. The energy transaction volume with the external market at time t, the energy allocated to energy providers in various industrial parks, and demand response information are used as the action space. The state at time t Transition to the state at time t+1 And the reward function is based on maximizing the operating revenue of the industrial park's energy operation and management provider. ; The electricity market clearing unit and the heat market clearing unit are defined as reinforcement learning agents: the energy price and power declared by the energy supplier at time t, and the energy price and power declared by the industrial park energy operator are used as the state space. The cleared electricity price, cleared heat price, and cleared quantity at time t are used as the action space. The state at time t Transition to the state at time t+1 and using the maximization of social welfare as the reward function. .

8. The multi-industrial park IES optimized operation method according to claim 1, characterized in that, In S5, the intelligent agents of the industrial park energy operation and management provider, the intelligent agents of the energy supplier, and the intelligent agents of the electricity market clearing unit and the heat market clearing unit conduct transactions, including: At the start of the initial bidding, the industrial park energy operation and management intelligent agent and the energy supplier intelligent agent respectively submit their respective energy prices and power to the electricity market clearing unit intelligent agent and the heat market clearing unit intelligent agent. Subsequently, the electricity market clearing unit intelligent agent and the heat market clearing unit intelligent agent process the submitted data according to the bidding and clearing mechanism and output clearing data information. The power market clearing unit and the heat market clearing unit are designated as leaders, while the industrial park energy operation and management unit and the energy supplier are designated as followers. The clearing data is analyzed based on game theory market interest equilibrium to determine whether direct clearing is possible. If interest is not balanced, an interest imbalance signal is sent to the industrial park energy operation and management unit and the energy supplier. The units then submit a new bid based on the clearing data until interest equilibrium is reached and the final clearing data is output. Transactions are conducted between the industrial park energy operation and management intelligent agent and the energy service intelligent agents of each industrial park, including: The intelligent agent of the energy operation and management company in the industrial park is regarded as the leader, and the intelligent agents of the energy companies in each industrial park are regarded as followers. The leader formulates pricing strategies based on the transaction information submitted by each follower and publishes them to each follower. Each follower adjusts its actual energy demand according to the price signal, resubmits transaction information, and constructs a master-slave game model and outputs the transaction information between the two after solving it. The transactions between the energy business intelligence agents in the various industrial parks include: defining the transactions between the energy business intelligence agents in the various industrial parks as cooperative game transactions; taking the minimization of the cost difference before and after the energy business intelligence agents in the various industrial parks participate in Nash bargaining as the objective; using the transaction electricity price between the energy business intelligence agents in the various industrial parks as the decision variable; constructing a cooperative game Nash bargaining model; and solving and outputting the transaction information between the energy business intelligence agents in the various industrial parks.

9. The multi-industrial park IES optimized operation method according to claim 1, characterized in that, The solution methods for the first-layer master-slave game optimization operation model, the second-layer master-slave game optimization operation model, and the third-layer cooperative game optimization operation model adopt deep reinforcement learning algorithms.

Images