CN116128543B - Comprehensive simulation operation method and system for load declaration and clearing of electricity selling company - Google Patents
Comprehensive simulation operation method and system for load declaration and clearing of electricity selling company Download PDFInfo
- Publication number
- CN116128543B CN116128543B CN202211622614.4A CN202211622614A CN116128543B CN 116128543 B CN116128543 B CN 116128543B CN 202211622614 A CN202211622614 A CN 202211622614A CN 116128543 B CN116128543 B CN 116128543B
- Authority
- CN
- China
- Prior art keywords
- load
- company
- electricity
- clearing
- selling company
- Prior art date
- Legal status (The legal status 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 status listed.)
- Active
Links
- 230000005611 electricity Effects 0.000 title claims abstract description 138
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000004088 simulation Methods 0.000 title claims abstract description 28
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 23
- 230000009471 action Effects 0.000 claims description 35
- 230000006870 function Effects 0.000 claims description 17
- 238000012549 training Methods 0.000 claims description 17
- 230000001186 cumulative effect Effects 0.000 claims description 11
- 238000010276 construction Methods 0.000 claims description 9
- 238000005457 optimization Methods 0.000 claims description 6
- 238000003860 storage Methods 0.000 claims description 4
- 239000003795 chemical substances by application Substances 0.000 description 29
- 230000006399 behavior Effects 0.000 description 10
- 238000012795 verification Methods 0.000 description 10
- 230000008569 process Effects 0.000 description 9
- 230000002787 reinforcement Effects 0.000 description 8
- 238000005070 sampling Methods 0.000 description 6
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 230000007246 mechanism Effects 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000033001 locomotion Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000010977 unit operation Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 239000011668 ascorbic acid Substances 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000002860 competitive effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012854 evaluation process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000000556 factor analysis Methods 0.000 description 1
- 238000013277 forecasting method Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000013486 operation strategy Methods 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- 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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06311—Scheduling, planning or task assignment for a person or group
-
- 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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0206—Price or cost determination based on market factors
-
- 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
-
- 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
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Accounting & Taxation (AREA)
- Human Resources & Organizations (AREA)
- Entrepreneurship & Innovation (AREA)
- Finance (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Educational Administration (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a comprehensive simulation operation method and system for load reporting and clearing of an electricity selling company, comprising the following steps: according to the acquired electric power market clearing information, a Markov decision model is built with the aim of minimizing the comprehensive cost of an electric power selling company; adopting a double-depth Q network algorithm to obtain an optimal strategy for reporting the load of an electricity selling company; and aiming at minimizing the running cost of the power system, taking system constraint, unit constraint and network security constraint into consideration, constructing a market clearing model, carrying out power market clearing simulation according to the load reporting optimal strategy of the electric company, and obtaining the power market clearing information of the next stage so as to formulate the load reporting optimal strategy of the electric company in the next stage. The load reporting decision-making behavior of the electricity selling company is intelligently simulated by adopting a double-depth Q network algorithm; by providing a market clearing model, a data-driven decision strategy is embedded in the market clearing model to reduce the clearing cost of an electricity company in market transactions.
Description
Technical Field
The invention relates to the technical field of electric power markets, in particular to a comprehensive simulation operation method and system for load reporting and clearing of an electricity selling company.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
The intelligent decision method is researched in the face of new situations of various electric power selling companies participating in electric power market transaction, and the intelligent decision method has very important significance and value for optimizing self behavior strategies of the electric power selling companies and realizing income maximization and optimal allocation of electric power market resources.
In the electric power market environment, an electricity selling company directly participates in market transaction, and electricity purchasing cost is reduced by changing self decision-making behavior; the main business is to purchase electric energy in each trade market and sell the electric energy to various end users in a retail mode. In which an electricity vending company typically provides a fixed price of electricity supply to the user, the rate adjustment is not frequent. In this context, in order to ensure self-income, the electricity selling company must minimize the cost of electricity purchasing on the basis of satisfying the user's demand, maintaining the average electricity purchasing price lower than the electricity selling price. Thus, the major challenges facing electricity companies are fluctuating electricity prices and random consumer load demands.
Many students at home and abroad research the electricity purchasing decision of electricity selling companies, and the system introduces the decision problem under the uncertainty of the electric power market. Research mainly comprises load prediction, electricity purchasing strategies, retail pricing strategies and the like. Load prediction is the basic basis for the electricity selling company to make optimal decisions; in the actual power market, however, the accuracy of load forecasting remains a difficult problem for power companies due to data quality limitations and the influence of many variables.
The electric energy purchasing process is a key step of decision making, and a plurality of research results exist in the related field at home and abroad, and the research results mainly focus on influence factor analysis and purchasing optimization strategy formulation in the purchasing process. But in different cases the above-described method lacks versatility.
Retail pricing strategies are one of the core challenges for retailers to increase profitability, but existing approaches require implementation of fixed models and lack the ability to accommodate complex and dynamic environments.
Disclosure of Invention
In order to solve the problems, the invention provides a comprehensive simulation operation method and a system for reporting and clearing the load of an electricity selling company, which adopts a double-depth Q network algorithm to intelligently simulate the load reporting decision-making behavior of the electricity selling company; by providing a market clearing model, a data-driven decision strategy is embedded in the market clearing model to reduce the clearing cost of an electricity company in market transactions.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
In a first aspect, the invention provides a comprehensive simulation operation method for reporting and clearing loads of an electricity selling company, which comprises the following steps:
According to the acquired electric power market clearing information, a Markov decision model is built with the aim of minimizing the comprehensive cost of the electric power selling company, and a state space, an action space and a cumulative rewarding function of the electric power selling company are formulated;
Generating an action value of an agent of the electric company by adopting a double-depth Q network algorithm for the Markov decision model, and obtaining an optimal strategy of load declaration of the electric company by iterative training until the accumulated rewards of the agent of the electric company are maximum;
aiming at minimizing the running cost of the power system, taking system constraint, unit constraint and network security constraint into consideration to construct a market clearing model;
And according to the optimal strategy of the load reporting of the electric company, adopting a market clearing model to perform electric power market clearing simulation, and obtaining electric power market clearing information of the next stage so as to formulate the optimal strategy of the load reporting of the electric company in the next stage.
As an alternative embodiment, the electric power market clearing information includes: current load prediction, day-ahead electricity prices, real-time electricity prices, and real-time load demands.
As an alternative embodiment, the objective function of the market clearing model is:
Wherein: lambda i,t is the unit marginal cost of the generator i; p i,t is the output of the unit i in the t period; The no-load cost, the starting cost and the stopping cost of the unit i are respectively represented; u i,t is the start-stop state of the unit; lambda j,t is the electricity price corresponding to the section t of the electricity selling company j; q j,t is the electricity purchase amount of electricity company j in the section t.
As an alternative embodiment, the system constraint includes a system active power balance constraint; the active power balance constraint of the system is as follows:
∑qj,t-∑pi,t＝0
wherein q j,t is the electricity purchased by electricity selling company j in the t section; p i,t is the output of the unit i in the t period.
Alternatively, the system constraint includes a system rotation reserve constraint; the system rotation reserve constraint is:
wherein, p i,t is the output of the unit i in the t period; u i,t is the start-stop state of the unit.
Alternatively, the jackpot function is:
Wherein: an optimal strategy obtained by the agent in the period t for the electricity selling company d; /(I) Comprehensive electricity price for market in the future; /(I)Comprehensive electricity price for real-time market; /(I)Obtaining a load prediction curve for an agent of an electricity selling company in a market in the day before; Is a real-time load curve; k d is the action value of the agent of the electricity selling company, and M is the uniform equal division times; d is an electricity selling company set; is a real-time load demand.
As an alternative embodiment, the state space is composed of a latest comprehensive price curve, a 96-point load prediction curve and a historical actual load curve; the action space is composed of action values generated by the agent of the electricity selling company.
In a second aspect, the present invention provides a comprehensive simulation operation system for reporting and clearing loads of an electricity selling company, including:
The decision model construction module is configured to construct a Markov decision model with the aim of minimizing the comprehensive cost of the electric company according to the acquired electric power market clearing information, and to formulate a state space, an action space and a cumulative rewarding function of the electric company;
The load reporting strategy optimization module is configured to adopt a double-depth Q network algorithm to the Markov decision model to generate an action value of an agent of the electricity selling company, and obtain an optimal load reporting strategy of the electricity selling company after iterative training until the accumulated rewards of the agent of the electricity selling company are maximum;
the system comprises a clearing model construction module, a model generation module and a model generation module, wherein the clearing model construction module is configured to take system constraint, unit constraint and network security constraint into consideration to construct a market clearing model aiming at minimizing the running cost of the power system;
The simulation module is configured to adopt a market clearing model to simulate the electric power market clearing according to the load reporting optimal strategy of the electric power selling company, and obtain electric power market clearing information of the next stage so as to formulate the load reporting optimal strategy of the electric power selling company in the next stage.
In a third aspect, the invention provides an electronic device comprising a memory and a processor and computer instructions stored on the memory and running on the processor, which when executed by the processor, perform the method of the first aspect.
In a fourth aspect, the present invention provides a computer readable storage medium storing computer instructions which, when executed by a processor, perform the method of the first aspect.
Compared with the prior art, the invention has the beneficial effects that:
the invention provides a comprehensive simulation operation method and a system for load declaration and clearing of an electricity selling company, which effectively apply deep reinforcement learning to simulate market quotation behaviors of the electricity selling company, and make Markov decision behaviors so as to form an intelligent quotation decision system of the electricity selling company; aiming at different types of market users, the decision-making model of the electricity selling company can stably promote the decision-making level; and developing a quick clearing model, and carrying out market clearing after the electric company obtains an optimal load reporting strategy, so as to compare and analyze the clearing cost per se and improve the use efficiency of a decision and clearing comprehensive system of the electric company.
The invention provides a comprehensive simulation operation method and a system for reporting and clearing the load of an electric company, which are used for formulating an action space, a state space and a cumulative rewarding function based on a Markov decision model, and iteratively simulating the load reporting decision behavior of an agent of the electric company by adopting an efficient and reliable double-depth Q network algorithm. The intelligent decision model of the electricity selling company based on the deep reinforcement learning has the characteristics of data driving, no model, closed-loop control and the like, can overcome the problems of complex calculation and expansibility related to an accurate model, and can also meet the real-time requirement of intelligent decision of the electricity selling company.
The invention provides a comprehensive simulation operation method and system for reporting and clearing the load of an electricity selling company, which aims at the electricity selling company in the market in the day-ahead, and for intelligently simulating the decision-making behavior of the load of the electricity selling company, adopts a double-depth Q network algorithm to simulate the decision-making behavior of the intelligent agent of the electricity selling company and provides a data-driven intelligent decision-making strategy; secondly, a market clearing model is provided, and a data-driven decision strategy is embedded in the market clearing model, so that the clearing cost of an electricity selling company in market transaction is reduced.
Additional aspects of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention.
FIG. 1 is a schematic diagram of an implementation mechanism of an intelligent decision model for reporting the load of an electricity-selling company provided in embodiment 1 of the present invention;
Fig. 2 is a schematic diagram of implementation of integrated decision making and clearing of a load declaration of an electric company based on deep reinforcement learning according to embodiment 1 of the present invention;
FIG. 3 is a block diagram of a comprehensive system for determining and clearing the load declaration of an electric company based on deep reinforcement learning according to embodiment 1 of the present invention;
FIG. 4 is a flowchart of a comprehensive simulation operation method for reporting and clearing the load of an electric company provided in embodiment 1 of the present invention;
FIG. 5 is a topology structure diagram of an IEEE 8 node system in example verification provided in embodiment 1 of the present invention;
FIG. 6 is a graph of the latest historical real-time load in the example verification provided in embodiment 1 of the present invention;
FIG. 7 is a graph of historical day-ahead, real-time electricity prices for example verification provided in example 1 of the present invention;
FIG. 8 is a graph of current load prediction in example verification provided in example 1 of the present invention;
FIG. 9 is a graph of the cumulative rewards and average rewards variation of the training process of the example verification provided in example 1 of the invention;
FIG. 10 is a graph showing the optimal decision of the electric company after training in the example verification provided in the embodiment 1 of the present invention;
FIG. 11 is a graph showing the comparison between the optimal reporting curve and the load prediction of the electric company 1 in the example verification provided in the embodiment 1 of the present invention;
FIG. 12 is a graph showing the comparison between the optimal reporting curve and the load forecast of the electric company 2 in the example verification provided in the embodiment 1 of the present invention;
FIG. 13 is a graph showing the comparison between the optimal reporting curve and the load forecast of the electric company 3 in the example verification provided in the embodiment 1 of the present invention;
Fig. 14 is a graph showing the ratio of the electricity purchase cost obtained by the electricity selling company according to the load prediction to the electricity purchase cost based on DDQN algorithm in the example verification provided in the embodiment 1 of the present invention.
Detailed Description
The invention is further described below with reference to the drawings and examples.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the invention. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular forms also are intended to include the plural forms, and furthermore, it is to be understood that the terms "comprises" and "comprising" and any variations thereof are intended to cover non-exclusive inclusions, such as, for example, processes, methods, systems, products or devices that comprise a series of steps or units, are not necessarily limited to those steps or units that are expressly listed, but may include other steps or units that are not expressly listed or inherent to such processes, methods, products or devices.
Embodiments of the invention and features of the embodiments may be combined with each other without conflict.
Example 1
The embodiment provides a comprehensive simulation operation method for decision-making and clearing of an electricity selling company based on deep reinforcement learning, which comprises the following steps:
According to the acquired electric power market clearing information, a Markov decision model is built with the aim of minimizing the comprehensive cost of the electric power selling company, and a state space, an action space and a cumulative rewarding function of the electric power selling company are formulated;
Generating an action value of an agent of the electric company by adopting a double-depth Q network algorithm for the Markov decision model, and obtaining an optimal strategy of load declaration of the electric company by iterative training until the accumulated rewards of the agent of the electric company are maximum;
aiming at minimizing the running cost of the power system, taking system constraint, unit constraint and network security constraint into consideration to construct a market clearing model;
And according to the optimal strategy of the load reporting of the electric company, adopting a market clearing model to perform electric power market clearing simulation, and obtaining electric power market clearing information of the next stage so as to formulate the optimal strategy of the load reporting of the electric company in the next stage.
The method of the present embodiment is described in detail below with reference to fig. 1-4.
In this embodiment, various types of clearing information sent by the electric power market include: current load prediction, day-ahead electricity price, real-time electricity price and real-time load demand;
according to the various clearing information, a Markov decision model is built with the aim of minimizing the comprehensive cost of the electric company, and an action space, a state space and a cumulative rewarding function of the electric company are formulated;
and dynamically generating an action value of the electric company agent by adopting a double-depth Q network (DDQN) algorithm on the Markov decision model, and iteratively training a load reporting decision curve until the accumulated rewards of the electric company agent are maximum, thereby obtaining an optimal electric company load reporting decision curve.
In this embodiment, in the market before date, the electricity selling company may make a load reporting optimal policy according to information such as load prediction, historical real-time load, current price before date, and real-time price of electricity. The electricity market can be seen as a competitive environment for electricity vendors, each of which is an agent that reacts to changes in the market environment. In fact, this is a typical Markov decision behavior, reinforcement Learning (RL) is an effective way to solve the Markov Decision Problem (MDP). According to the characteristics of the electricity selling company load reporting decision, an MDP model is designed as follows:
state space: the state space of the MDP consists of a latest comprehensive price curve issued by a power dispatching mechanism, a 96-point load prediction curve and a historical actual load curve, as shown in a formula (1).
Action space: the action value is generated by the agent of the electricity selling company d, and the action space is expressed as follows:
A＝{kd,min,...,kd,...,kd,max}|kd＝(kd,max-kd,min)/M (2)
Bonus function: in the power market environment, the optimal control aim is to minimize the long-term electricity purchasing cost of an electricity selling company. In the current work, delay rewards are adopted as the optimization basis of the controller in consideration of the time sequence of market release information.
The utility company declares the load curve in the future, i.e. the work order output by the controller will instruct the utility company to declare the load in the future one hour. Because the release information of the market stage before the day is limited, the state space only has the predicted load of the market before the day and the comprehensive electricity price before the day in the load reporting process, the market in the stage needs to combine the latest information such as the real-time load, the real-time comprehensive electricity price and the like released by the market in the last period to develop the settlement of the electricity cost of the market reporting.
Based on the above flow, the reward function of each action of the controller is:
Wherein: Comprehensive electricity price for market in the future; /(I) Comprehensive electricity price for real-time market; /(I)Obtaining a load prediction curve for an agent of an electricity selling company in a market in the day before; /(I)Is a real-time load curve; k d is the action value of the agent of the electricity selling company, and M is the uniform equal division times; /(I)An optimal strategy obtained by the agent in the period t for the electricity selling company d; d is an electricity selling company set; is a real-time load demand.
In the embodiment, a DDQN algorithm architecture decision model is adopted, a Markov decision model is utilized to determine a learning method of an optimal strategy, and the optimal strategy is declared by simulating the load of an electricity selling company according to the learning method of the optimal strategy. The DDQN algorithm is a deep reinforcement learning method based on Q learning and DQN, not only inherits the value-based characteristic of the Q learning, but also introduces a neural network to predict the value of the behavior so as to avoid the problem that the Q learning is easy to occur. In addition, DDQN decouples the selection process and the evaluation process, avoiding the problem of overestimation of the Q value. The DDQN algorithm updates the agent parameters during training by the prize values returned by the external environment. The agent dynamically selects the action value through different state spaces to obtain the highest prize.
In fact, the DDQN algorithm differs from the DQN algorithm in the process of selecting the Q value. DQN always selects the maximum Q value of the target network; and DDQN first selects an action with the maximum Q value from the online network using equation (4), then evaluates the action policy by calculating the corresponding target Q value through the target network, and evaluates by equation (5), the time difference error (TD error) of DDQN is expressed as equation (6).
a*＝argmaxaQ(St+1,a;θt) (4)
Wherein: q (S t+1,a;θt) is the current Q value of action a t in state S t+1 output by the online Q network; q (S t+1,argmaxaQ(St+1,a;θt);θ′t) is a target Q value output by a target Q network; θ t and θ' t are online Q network and target Q network parameters, respectively; gamma is the attenuation coefficient; r t+1 is the jackpot value.
In DDQN algorithm, in order to break the association relation between sample data, probability extraction sample learning such as experience playback pool is adopted to update neural network parameters. However, for the sparse rewards, there are fewer samples that can excite the controller to act correctly, resulting in a less efficient training of the uniform sampling mechanism. Therefore, when the experience playback pool sample data is extracted for training, the sampling probability of each experience sample is determined according to the quality of the experience sample by adopting a priority sampling mode, so that samples with higher priorities are extracted and learned more frequently. The sampling mode can obviously improve the convergence speed of DDQN algorithm and reduce the number of samples required by the convergence of the neural network. The above procedure can be defined as formula (7):
wherein: p (i) is the conversion priority at sample i, with an exponential power distribution characteristic, the exponent determining the priority; Can be obtained from the absolute value of the TD error; p i can be expressed as p i =1/rank (i), and rank (i) is the rank of the conversion.
When using probability distribution sampling for preferential playback, there is a bias in the estimation of the expected value. To correct for this deviation, it is typically multiplied by an importance sampling weight, as shown in equation (8):
wherein: n is a capacity replay buffer; beta is an index value; δ j+1 is obtained by formula (6); η is the step size.
In the embodiment, constraint conditions of system balance, power grid safety and unit operation are considered, a market clearing model is established, and a decision model for reporting the load of the electricity selling company is embedded in the market clearing model so as to simulate the market clearing of an optimal strategy for reporting the load of the electricity selling company;
According to the optimal strategy of the load declaration of the electric company, carrying out electric power market clearing simulation through a market clearing model to obtain the current clearing cost of the electric company for comparison and evaluation; meanwhile, clear information of the electric power market in the next stage is formed, so that an electric power selling company can formulate a load reporting optimal strategy in the next stage; the method is used for realizing balance of market supply and demand, meeting load demands of users, and simultaneously controlling power grid operation strategies such as power output of a generator set, start and stop of loads and the like in a power grid while meeting the load demands of the users, so as to realize balance of power supply demands of the power grid and power consumption demands of the users.
In the embodiment, the intelligent decision model for reporting the load of the electric company based on deep reinforcement learning adjusts the action value of the intelligent agent when meeting the constraint condition of the reporting load of the electric company, so that the cost of the electric company is minimized; meanwhile, a market clearing model aiming at minimizing the running cost of the power system is adopted, and in a certain scheduling period, the given load balance is met, meanwhile, certain boundary conditions and standby requirements are met, and the on-off state of the unit is reasonably distributed, so that the power generation cost is minimized.
In the market in the day-ahead, an independent operator firstly distributes related information of market in the day-ahead and real-time market transaction to each electricity selling company; secondly, the electricity selling company optimizes the strategy of the electricity selling company through an intelligent decision model and forms a declaration load curve. And then, calling a market clearing program to calculate the transaction clearing result of the declared strategy, and comparing and evaluating the transaction clearing result with the transaction clearing result obtained by the forecasting method.
In the embodiment, a multi-type power supply is connected into a power system, constraint conditions such as system balance, power grid safety and unit operation are considered, and a market clearing model aiming at minimizing the operation cost of the power system is established; wherein the objective function is expressed as:
Wherein: lambda i,t is the unit marginal cost of the generator i; p i,t is the output of the unit i in the t period; The no-load cost, the starting cost and the stopping cost of the unit i are respectively represented; u i,t is the start-stop state of the unit; lambda j,t is the electricity price corresponding to the section t of the electricity selling company j; q j,t is the electricity purchase amount of electricity company j in the section t.
In this embodiment, constraint conditions of the market clearing model include system constraint, unit constraint, and network security constraint;
Wherein the system constraint comprises a system active power balance constraint and a system rotation reserve constraint;
(1) The active power balance constraint of the system is as follows:
Σqj,t-Σpi,t＝0 (11)
(2) The system rotation reserve constraint is:
the unit constraint is as follows:
wherein: i=1, 2,..n.
The network security constraints are:
Fl-Fl M≤0 (15)
-Fl-Fl M≤0 (16)
Wherein: f l is the power flow of line l; f l M is line l power limit.
In this embodiment, the market clearing model adopts mixed integer programming modeling to solve, performs electric power market clearing simulation according to the load reporting optimal strategy of the electric power selling company, and obtains electric power market clearing information of the next stage so as to formulate the load reporting optimal strategy of the electric power selling company in the next stage.
As shown in fig. 5, the present embodiment introduces an IEEE 8 node system as an example, verifying the feasibility of the method of the present embodiment. Wherein, the transmission capacity of each line is set to 300MW, the relevant parameters of the generator set are shown in table 1, and the detailed information of DDQN super parameters is listed in table 2. In the action space, the capacity of each electric company agent is set to k min to-50 MW and k max to 50MW. The motion space is divided into 20 motions at equal intervals, that is, M is 20. The historical real-time load of the electricity selling company is shown in fig. 6, the day-ahead electricity price and the real-time electricity price are shown in fig. 7, and the predicted load curve of the electricity selling company is shown in fig. 8.
Table 1 generator set parameters related to the generator set
Table 2 DDQN super parameters
| Super parameter | Value of |
| E | 300 |
| T | 96 |
| Batch size | 32 |
| Experience pool size | 5000 |
| Attenuation Rate | 0.9 |
| K | 50 |
| α | 0.6 |
| β | 0.4 |
| Learning rate | 0.001 |
In each iteration, the agent adjusts the updating of the parameters according to the rewarding value returned by the external environment in the training process, and obtains better actions from the action space. Fig. 9 shows that the prize value increases with the number of exercises, indicating that the agent has learned effectively from previous experience. It can be seen that during the sample accumulation phase, the evaluation of the Q value and training of the neural network is not ideal due to the insufficient number of samples stored in the experience pool, and thus the revenue of this phase is low. As samples accumulate, agent learning is continually optimized, so revenue also rises significantly, and eventually reaches a steady value as training times increase. This means that the parameters of the agent have converged and can be applied in the actual scenario. Fig. 10 shows the optimal decision curve of the electricity vending company.
In an actual application scenario, the data sample of the training set can be used as a test set, and the 96-point test set in the training steady state is intercepted as an optimal strategy of an electricity selling company. In order to highlight the optimization effect, the DDQN obtained 96-point optimal decision curve is compared with a load prediction curve, as shown in fig. 11-13; meanwhile, fig. 14 shows the ratio of the electricity purchase cost obtained by the electricity selling company according to the load prediction to the electricity purchase cost based on DDQN algorithm. It can be observed that the ratio is almost greater than 1, and it can be stated that the electricity purchase cost obtained by using DDQN algorithm is generally smaller than that obtained using the predictive method. Therefore, the method can significantly improve the decision level of the electricity selling company.
To further verify the performance of the proposed method, the present embodiment performs simulation in the following two scenarios, as follows: (1) The electricity vending company uses the proposed method to optimize its own strategy and then makes market clearing. (2) The electricity selling company runs the market clear according to the load prediction curve issued by the ISO. The optimal decision curves and load prediction curves of the electricity vending company are shown in fig. 11-13. The results of the clearing of the electricity selling company in both cases are shown in table 3, and it is apparent that the electric clearing cost of the electricity selling company obtained by DDQN method is lower than the clearing cost of the use prediction method.
TABLE 3 analysis of market clearing results
Example 2
The embodiment provides a comprehensive simulation operation system for reporting and clearing loads of an electricity selling company, which comprises the following components:
The decision model construction module is configured to construct a Markov decision model with the aim of minimizing the comprehensive cost of the electric company according to the acquired electric power market clearing information, and to formulate a state space, an action space and a cumulative rewarding function of the electric company;
The load reporting strategy optimization module is configured to adopt a double-depth Q network algorithm to the Markov decision model to generate an action value of an agent of the electricity selling company, and obtain an optimal load reporting strategy of the electricity selling company after iterative training until the accumulated rewards of the agent of the electricity selling company are maximum;
the system comprises a clearing model construction module, a model generation module and a model generation module, wherein the clearing model construction module is configured to take system constraint, unit constraint and network security constraint into consideration to construct a market clearing model aiming at minimizing the running cost of the power system;
The simulation module is configured to adopt a market clearing model to simulate the electric power market clearing according to the load reporting optimal strategy of the electric power selling company, and obtain electric power market clearing information of the next stage so as to formulate the load reporting optimal strategy of the electric power selling company in the next stage.
It should be noted that the above modules correspond to the steps described in embodiment 1, and the above modules are the same as examples and application scenarios implemented by the corresponding steps, but are not limited to those disclosed in embodiment 1. It should be noted that the modules described above may be implemented as part of a system in a computer system, such as a set of computer-executable instructions.
In further embodiments, there is also provided:
An electronic device comprising a memory and a processor and computer instructions stored on the memory and running on the processor, which when executed by the processor, perform the method described in embodiment 1. For brevity, the description is omitted here.
It should be understood that in this embodiment, the processor may be a central processing unit CPU, and the processor may also be other general purpose processors, digital signal processors DSP, application specific integrated circuits ASIC, off-the-shelf programmable gate array FPGA or other programmable logic device, discrete gate or transistor logic devices, discrete hardware components, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The memory may include read only memory and random access memory and provide instructions and data to the processor, and a portion of the memory may also include non-volatile random access memory. For example, the memory may also store information of the device type.
A computer readable storage medium storing computer instructions which, when executed by a processor, perform the method described in embodiment 1.
The method in embodiment 1 may be directly embodied as a hardware processor executing or executed with a combination of hardware and software modules in the processor. The software modules may be located in a random access memory, flash memory, read only memory, programmable read only memory, or electrically erasable programmable memory, registers, etc. as well known in the art. The storage medium is located in a memory, and the processor reads the information in the memory and, in combination with its hardware, performs the steps of the above method. To avoid repetition, a detailed description is not provided herein.
Those of ordinary skill in the art will appreciate that the elements of the various examples described in connection with the present embodiments, i.e., the algorithm steps, can be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
While the foregoing description of the embodiments of the present invention has been presented in conjunction with the drawings, it should be understood that it is not intended to limit the scope of the invention, but rather, it is intended to cover all modifications or variations within the scope of the invention as defined by the claims of the present invention.
Claims (8)
1. The utility model provides a comprehensive simulation operation method for the load declaration and clearing of an electricity selling company, which is characterized by comprising the following steps:
According to the acquired electric power market clearing information, a Markov decision model is built with the aim of minimizing the comprehensive cost of the electric power selling company, and a state space, an action space and a cumulative rewarding function of the electric power selling company are formulated;
Generating an action value of an agent of the electric company by adopting a double-depth Q network algorithm for the Markov decision model, and obtaining an optimal strategy of load declaration of the electric company by iterative training until the accumulated rewards of the agent of the electric company are maximum;
aiming at minimizing the running cost of the power system, taking system constraint, unit constraint and network security constraint into consideration to construct a market clearing model;
According to the optimal strategy of the load declaration of the electric company, adopting a market clearing model to perform electric power market clearing simulation, and obtaining electric power market clearing information of the next stage so as to formulate the optimal strategy of the load declaration of the electric company in the next stage;
the objective function of the market clearing model is as follows:
Wherein: lambda i,t is the unit marginal cost of the generator i; p i,t is the output of the unit i in the t period; The no-load cost, the starting cost and the stopping cost of the unit i are respectively represented; u i,t is the start-stop state of the unit; lambda j,t is the electricity price corresponding to the section t of the electricity selling company j; q j,t is the electricity quantity purchased by the electricity selling company j in the section t;
The cumulative bonus function is:
Wherein: an optimal strategy obtained by the agent in the period t for the electricity selling company d; /(I) Comprehensive electricity price for market in the future; /(I)Comprehensive electricity price for real-time market; /(I)Obtaining a load prediction curve for an agent of an electricity selling company in a market in the day before; /(I)Is a real-time load curve; k d is the action value of the agent of the electricity selling company, and M is the uniform equal division times; d is an electricity selling company set; /(I)Is a real-time load demand.
2. The integrated simulation operation method for load declaration and clearing of an electric company according to claim 1, wherein the electric power market clearing information includes: current load prediction, day-ahead electricity prices, real-time electricity prices, and real-time load demands.
3. The method for integrated simulated operation of utility load reporting and clearing as claimed in claim 1, wherein said system constraints comprise system active power balance constraints; the active power balance constraint of the system is as follows:
∑qj,t-∑pi,t＝0
wherein q j,t is the electricity purchased by electricity selling company j in the t section; p i,t is the output of the unit i in the t period.
4. The method for integrated simulation operation of load declaration and clearing of an electric company according to claim 1, wherein the system constraint includes a system rotation reserve constraint; the system rotation reserve constraint is:
wherein, p i,t is the output of the unit i in the t period; u i,t is the start-stop state of the unit.
5. The integrated simulation operation method for the load declaration and clearing of an electric company according to claim 1, wherein the state space consists of a latest integrated price curve, a 96-point load prediction curve and a historical actual load curve; the action space is composed of action values generated by the agent of the electricity selling company.
6. The utility model provides a comprehensive simulation operation system of selling electric company load declaration and clearing which characterized in that includes:
The decision model construction module is configured to construct a Markov decision model with the aim of minimizing the comprehensive cost of the electric company according to the acquired electric power market clearing information, and to formulate a state space, an action space and a cumulative rewarding function of the electric company;
The load reporting strategy optimization module is configured to adopt a double-depth Q network algorithm to the Markov decision model to generate an action value of an agent of the electricity selling company, and obtain an optimal load reporting strategy of the electricity selling company after iterative training until the accumulated rewards of the agent of the electricity selling company are maximum;
the system comprises a clearing model construction module, a model generation module and a model generation module, wherein the clearing model construction module is configured to take system constraint, unit constraint and network security constraint into consideration to construct a market clearing model aiming at minimizing the running cost of the power system;
The simulation module is configured to adopt a market clearing model to simulate the electric power market clearing according to the load reporting optimal strategy of the electric power selling company, and obtain electric power market clearing information of the next stage so as to formulate the load reporting optimal strategy of the electric power selling company in the next stage;
the objective function of the market clearing model is as follows:
Wherein: lambda i,t is the unit marginal cost of the generator i; p i,t is the output of the unit i in the t period; The no-load cost, the starting cost and the stopping cost of the unit i are respectively represented; u i,t is the start-stop state of the unit; lambda j,t is the electricity price corresponding to the section t of the electricity selling company j; q j,t is the electricity quantity purchased by the electricity selling company j in the section t;
The cumulative bonus function is:
Wherein: an optimal strategy obtained by the agent in the period t for the electricity selling company d; /(I) Comprehensive electricity price for market in the future; /(I)Comprehensive electricity price for real-time market; /(I)Obtaining a load prediction curve for an agent of an electricity selling company in a market in the day before; /(I)Is a real-time load curve; k d is the action value of the agent of the electricity selling company, and M is the uniform equal division times; d is an electricity selling company set; /(I)Is a real-time load demand.
7. An electronic device comprising a memory and a processor and computer instructions stored on the memory and running on the processor, which when executed by the processor, perform the method of any one of claims 1-5.
8. A computer readable storage medium storing computer instructions which, when executed by a processor, perform the method of any of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211622614.4A CN116128543B (en) | 2022-12-16 | 2022-12-16 | Comprehensive simulation operation method and system for load declaration and clearing of electricity selling company |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211622614.4A CN116128543B (en) | 2022-12-16 | 2022-12-16 | Comprehensive simulation operation method and system for load declaration and clearing of electricity selling company |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116128543A CN116128543A (en) | 2023-05-16 |
| CN116128543B true CN116128543B (en) | 2024-05-24 |
Family
ID=86305596
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211622614.4A Active CN116128543B (en) | 2022-12-16 | 2022-12-16 | Comprehensive simulation operation method and system for load declaration and clearing of electricity selling company |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116128543B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117875153A (en) * | 2023-05-23 | 2024-04-12 | 海南晴杉慧源科技有限公司 | Electric power market simulation system, method, equipment and storage medium |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108711890A (en) * | 2018-06-27 | 2018-10-26 | 广东电网有限责任公司 | Ahead market goes out clearing method, system, device and computer readable storage medium |
| KR20190132193A (en) * | 2018-05-18 | 2019-11-27 | 한양대학교 에리카산학협력단 | A Dynamic Pricing Demand Response Method and System for Smart Grid Systems |
| CN110909910A (en) * | 2019-09-18 | 2020-03-24 | 浙江大学 | Novel deviation electric quantity checking mechanism optimization design method based on PBR |
| CN111476399A (en) * | 2020-03-08 | 2020-07-31 | 云南电网有限责任公司 | Electricity purchasing method and system for electricity selling company under spot transaction mechanism |
| CN111859653A (en) * | 2020-07-13 | 2020-10-30 | 中国电力科学研究院有限公司 | Method and system for simulating market behavior of power selling company in direct transaction of power consumer |
| CN112446540A (en) * | 2020-11-25 | 2021-03-05 | 广东电网有限责任公司电力调度控制中心 | Electric power spot market clearing and settlement optimizing method and device |
| CN112651770A (en) * | 2020-12-07 | 2021-04-13 | 山东大学 | Load declaration optimization method and system for power selling merchants in power spot market |
| CN113032995A (en) * | 2021-03-23 | 2021-06-25 | 山东大学 | Electric power market simulation optimization and operation method, system and simulation platform |
| CN113326994A (en) * | 2021-07-06 | 2021-08-31 | 华北电力大学 | Virtual power plant energy collaborative optimization method considering source load storage interaction |
| CN114897552A (en) * | 2022-03-18 | 2022-08-12 | 南京工程学院 | Reinforced learning-based price reporting method for spot-market electricity selling merchants |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020143683A1 (en) * | 2000-02-24 | 2002-10-03 | Richard Taylor | Stockholder bonus and participation system |
| US20200327411A1 (en) * | 2019-04-14 | 2020-10-15 | Di Shi | Systems and Method on Deriving Real-time Coordinated Voltage Control Strategies Using Deep Reinforcement Learning |
| US11476669B2 (en) * | 2019-06-11 | 2022-10-18 | Tata Consultancy Services Limited | Method and system for building reinforcement learning (RL) based model for generating bids |
| TWI763087B (en) * | 2020-10-21 | 2022-05-01 | 國立清華大學 | Method and apparatus for peer-to-peer energy sharing based on reinforcement learning |
| US11887140B2 (en) * | 2021-02-25 | 2024-01-30 | Mitsubishi Electric Research Laboratories, Inc. | Risk-constrained optimization of virtual power plants in pool and future markets |
-
2022
- 2022-12-16 CN CN202211622614.4A patent/CN116128543B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20190132193A (en) * | 2018-05-18 | 2019-11-27 | 한양대학교 에리카산학협력단 | A Dynamic Pricing Demand Response Method and System for Smart Grid Systems |
| CN108711890A (en) * | 2018-06-27 | 2018-10-26 | 广东电网有限责任公司 | Ahead market goes out clearing method, system, device and computer readable storage medium |
| CN110909910A (en) * | 2019-09-18 | 2020-03-24 | 浙江大学 | Novel deviation electric quantity checking mechanism optimization design method based on PBR |
| CN111476399A (en) * | 2020-03-08 | 2020-07-31 | 云南电网有限责任公司 | Electricity purchasing method and system for electricity selling company under spot transaction mechanism |
| CN111859653A (en) * | 2020-07-13 | 2020-10-30 | 中国电力科学研究院有限公司 | Method and system for simulating market behavior of power selling company in direct transaction of power consumer |
| CN112446540A (en) * | 2020-11-25 | 2021-03-05 | 广东电网有限责任公司电力调度控制中心 | Electric power spot market clearing and settlement optimizing method and device |
| CN112651770A (en) * | 2020-12-07 | 2021-04-13 | 山东大学 | Load declaration optimization method and system for power selling merchants in power spot market |
| CN113032995A (en) * | 2021-03-23 | 2021-06-25 | 山东大学 | Electric power market simulation optimization and operation method, system and simulation platform |
| CN113326994A (en) * | 2021-07-06 | 2021-08-31 | 华北电力大学 | Virtual power plant energy collaborative optimization method considering source load storage interaction |
| CN114897552A (en) * | 2022-03-18 | 2022-08-12 | 南京工程学院 | Reinforced learning-based price reporting method for spot-market electricity selling merchants |
Non-Patent Citations (10)
| Title |
|---|
| A Deep Q Network Approach for Optimizing Offering Strategies in Electricity Markets;Yujian Ye等;IEEE;20190926;全文 * |
| Lu Kuan ; 等.Short-term electricity load forecasting method based on multilayered self-normalizing GRU network.IEEE.2018,全文. * |
| 不完全信息下基于多代理深度确定策略梯度算法 的发电商竞价策略;员江洋等;电网技术;20221231;第46卷(第12期);4832-4842 * |
| 任艺 ; 周明 ; 李庚银 ; .考虑用户需求响应的售电公司购售电决策双层模型.电力系统自动化.2017,(第14期),全文. * |
| 员江洋等.不完全信息下基于多代理深度确定策略梯度算法的发电商竞价策略.电网技术.2022,第46卷(第12期),4832-4842. * |
| 基于信息间隙决策理论的多种零售合同模式下售电公司购售电策略;唐力;刘继春;杨阳方;周玲;;电网技术;20190416(第06期);全文 * |
| 杨明等.Interval Estimation for Conditional Failure Rates of Transmission Lines With Limited Samples.IEEE TRANSACTIONS ON SMART GRID.2016,第9卷(第4期),2752-2763. * |
| 深度强化学习在电力市场参与者策略优化中的应用研究;高源;中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑;20211215(第12期);C042-50 * |
| 考虑可中断负荷的配售电公司最优购售电策略;邹鹏;李春晖;郭静;邢亚虹;陈启鑫;;南方电网技术(第02期);全文 * |
| 集成竞价仿真与告警管理的售电运营平台设计;蒋衍君;李鹏;孙常浩;;计算机时代;20180315(第03期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116128543A (en) | 2023-05-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Weitzel et al. | Energy management for stationary electric energy storage systems: A systematic literature review | |
| Wang et al. | Analysis of energy storage demand for peak shaving and frequency regulation of power systems with high penetration of renewable energy | |
| Gonzalez et al. | Forecasting power prices using a hybrid fundamental-econometric model | |
| Löhndorf et al. | Optimal day-ahead trading and storage of renewable energies—an approximate dynamic programming approach | |
| Zhang et al. | Bi-level stochastic real-time pricing model in multi-energy generation system: A reinforcement learning approach | |
| Jiang et al. | Research on short-term optimal scheduling of hydro-wind-solar multi-energy power system based on deep reinforcement learning | |
| CN112952831B (en) | Daily optimization operation strategy for providing stacking service by load side energy storage | |
| Zhao et al. | Storage or no storage: Duopoly competition between renewable energy suppliers in a local energy market | |
| CN116128543B (en) | Comprehensive simulation operation method and system for load declaration and clearing of electricity selling company | |
| CN112308411B (en) | Comprehensive energy station random planning method and system based on dynamic carbon trading model | |
| Wehinger et al. | Modeling electricity wholesale markets with model predictive and profit maximizing agents | |
| Zou et al. | Day-ahead energy sharing schedule for the P2P prosumer community using LSTM and swarm intelligence | |
| CN117578409A (en) | Multi-energy complementary optimization scheduling method and system in power market environment | |
| CN117726143A (en) | Environment-friendly micro-grid optimal scheduling method and system based on deep reinforcement learning | |
| US11221366B2 (en) | System and method for management of electrochemical energy storage devices | |
| CN114169916A (en) | Market member quotation strategy making method suitable for novel power system | |
| CN111126709B (en) | Method, device, computer equipment and storage medium for determining dispatch output | |
| Lau et al. | Strategy and modeling for building DR optimization | |
| Dokuchaev et al. | Optimal energy storing and selling in continuous time stochastic multi-battery setting | |
| Rasheed et al. | Artificial intelligence-enabled probabilistic load demand scheduling with dynamic pricing involving renewable resource | |
| Deng et al. | Energy Management of Price-Maker Community Energy Storage by Stochastic Dynamic Programming | |
| Zahid et al. | Analyses on the impact of consumers’ participation by demand response for flexibility procurement | |
| Ordoñez et al. | Comprehensive analysis of MPC-based energy management strategies for isolated microgrids empowered by storage units and renewable energy sources | |
| CN112184294A (en) | Practical reinforced learning automata method for optimizing quotation of limited information generator | |
| Cao et al. | LSTM-Aided Reinforcement Learning for Energy Management in Microgrid with Energy Storage and EV Charging |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |