CN105391090A - Multi-intelligent-agent multi-target consistency optimization method of intelligent power grid - Google Patents

Multi-intelligent-agent multi-target consistency optimization method of intelligent power grid Download PDF

Info

Publication number
CN105391090A
CN105391090A CN201510759625.0A CN201510759625A CN105391090A CN 105391090 A CN105391090 A CN 105391090A CN 201510759625 A CN201510759625 A CN 201510759625A CN 105391090 A CN105391090 A CN 105391090A
Authority
CN
China
Prior art keywords
load
netlogo
target
matlab
electricity price
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.)
Granted
Application number
CN201510759625.0A
Other languages
Chinese (zh)
Other versions
CN105391090B (en
Inventor
方周
付蓉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing Post and Telecommunication University
Nanjing University of Posts and Telecommunications
Original Assignee
Nanjing Post and Telecommunication University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing Post and Telecommunication University filed Critical Nanjing Post and Telecommunication University
Priority to CN201510759625.0A priority Critical patent/CN105391090B/en
Publication of CN105391090A publication Critical patent/CN105391090A/en
Application granted granted Critical
Publication of CN105391090B publication Critical patent/CN105391090B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks
    • H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
    • H02J3/46—Controlling of the sharing of output between the generators, converters, or transformers
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
    • H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]

Landscapes

  • Engineering & Computer Science (AREA)
  • Power Engineering (AREA)
  • Supply And Distribution Of Alternating Current (AREA)
  • Pharmaceuticals Containing Other Organic And Inorganic Compounds (AREA)

Abstract

The invention discloses a multi-intelligent-agent multi-target consistency optimization method of an intelligent power grid. The method is characterized in that different typical features of different power elements as well as objectives and requirements provided by the power elements respectively are analyzed according to characteristics of the multi-intelligent-agent cooperative control factor system of the intelligent power grid; when objectives are diversified, a proper objective function is selected to obtain an optimized operation control way and parameters, thereby guaranteeing reliability and economy during the system operation; and effectiveness of the optimization operation strategy is verified. According to the invention, a multi-intelligent-agent optimization model is established according to characteristics of the power grid and the load; a distributed output optimization algorithm with partial information sharing is studied and considered by using a multi-intelligent-agent theory; convergence of the algorithm can be analyzed based on different communication topology units; and simulated analyses are carried out on examples and correlated technologies for improving the convergence of the distributed algorithm are studied.

Description

A kind of intelligent grid multiple agent multiple target consistency optimization method
Technical field
The invention belongs to intelligent grid and optimize coordinated scheduling technical field, relate to the intelligent grid Optimal Operation Strategies of the multi-objective coordinated control of a kind of multiple agent, be specifically related to a kind of intelligent grid multiple agent multiple target consistency optimization method.
Background technology
Intelligent grid is an important branch of artificial intelligence, is the front subject of artificial intelligence in the world at the beginning of 20 end of the centurys to 21 century.Fast development along with computer technology, artificial intelligence theory, control theory and the continuous exploration to modern science, intelligent grid has become one of hot issue of different ambit research.The distributed collaboration of intelligent grid controls raising distribution network reliability, improve the quality of power supply, improve power distribution network performance driving economy, optimize power distribution network runs and all tools such as to arrange be of great significance.
Power-balance controls, i.e. Real-time Economic Dispatch is a basic problem in power system operation, and it refers to that generator and flexible load are under the condition meeting a series of operation constraint, make the optimization problem of the maximization of economic benefit of whole power system operation.Centralized optimization technology is adopted to solve Economic Dispatch Problem, comprising classic optimization method and modern artificial intelligence approach traditionally.
But when adopting centralized optimization method, system needs all generators and flexible load in control centre's issuing command scheduling whole system, and control centre needs to carry out information interaction with each scheduler object.Further, " plug and play " technology that the extensive infiltration of flexible load and force device need will make power network and communication network topological structure changeable, cause the communication topology construction cost that centralized optimization method needs are higher.Therefore, needing the optimized algorithm that adaptability is stronger, still can effectively run when losing efficacy in the limited and unreliable even control centre that communicates.
Summary of the invention
The object of the invention is to overcome the deficiencies in the prior art part, a kind of intelligent grid multiple agent multiple target consistency optimization method is provided, according to the interactive requirement of net lotus, in conjunction with dissimilar intelligent grid multiple agent characteristic, set up cooperation control model from the conforming angle of intelligent grid multiple agent multiple-target system.The present invention can not only according to power network and part throttle characteristics, set up the Optimized model of multiple agent, multi-agent theory research is utilized to consider the distributed generating optimization algorithm of part information sharing, according to the convergence of different communication topology parsers, simulation analysis can also be carried out to example and the constringent correlation technique of research raising distributed algorithm.
For solving the problems of the technologies described above, the present invention by the following technical solutions:
A kind of intelligent grid multiple agent multiple target consistency optimization method, it is characterized in that, according to the feature of the system of intelligent grid multiple agent Collaborative Control factor, analyze the different characteristic features of different force device, and respective proposed target call, choose appropriate target function when target is various and obtain optimizing operation control mode and parameter, reliability when guaranteeing system cloud gray model and economy, and verify the validity of Optimal Operation Strategies; Implementation step comprises:
Step 1, according to power system network structure, set up the union simulation platform based on MATLAB and NETLOGO, wherein, power system component model is set up in MATLAB, in NETLOGO, definition represents the intelligent body general module of power system component, and meanwhile, the data exchange interface module of building between MATLAB and NETLOGO realizes information interaction;
Step 2, for various load type, respectively according to load datum quantity, electricity price, and the goal orientation degree of each target of corresponding load, set up the load-Respondence to the Price of Electric Power characteristic model corresponding respectively to various load and power supply type; Described load comprises rigidity load and flexible load, and described power supply comprises distributed power source and energy-storage travelling wave tube; Wherein, rigidity load refers to the load not participating in electrical network interaction, and flexible load refers to the load participating in electrical network interaction;
Step 3, according to the load-Respondence to the Price of Electric Power characteristic model of the corresponding various load type of the difference set up in described step 2, obtains the target function of each target of each load respectively; And respectively for each load, the target function of each target of load is weighted process, obtains the general objective function of each load corresponding respectively;
Step 4, is randomly dispersed in each described load in the three-dimensional aspect of NETLOGO, obtains the initial policy of each load; For the network node in the three-dimensional aspect of NETLOGO, set electricity price at random, and set up load agency;
Step 5, using the initial policy of each load described as load datum quantity, respectively for the goal orientation degree of each target of each load, adopt the mode of+i or-i to obtain strategy corresponding to each load respectively, and form the set of strategies of each load in conjunction with the initial policy of each load; Described i is each step iteration step length, and described step refers to that electricity price often changes once, and the tactful respective change of load once;
Step 6, adopt the intelligent grid consistency optimized algorithm of the multi-objective coordinated control of multiple agent, respectively coordination computing is optimized to the general objective function of each load, and selects the strategy obtaining each load its maximum general objective functional value corresponding respectively, as the preference policy of each load;
Make x irepresent the state of force device, according to consistency protocol, when the state value of and if only if network opens up all nodes of bowl spares is all equal, the node of this network all reaches unanimously, that is:
x 1＝x 2＝L＝x n
Step 7, respectively according to the goal orientation degree of each target in the preference policy of each described load, moves to each load respectively in the three-dimensional aspect of NETLOGO on corresponding position, and upgrades the goal orientation degree of each target of each load; Then according to the load-Respondence to the Price of Electric Power characteristic model of correspondence, obtain the power of now each load, and in conjunction with the administration of load agency for corresponding load, obtain the gross power of each load agency respectively;
Step 8, the gross power that each described load is acted on behalf of is sent in MATLAB by NETLOGO, in MATLAB, obtain the electricity price of generator output and each network node corresponding, and be back in NETLOGO, upgrade the electricity price on map network node in the three-dimensional aspect of NETLOGO;
Step 9, using the electricity price on each network node in three-dimensional for described NETLOGO aspect as traction signal, and is acted on behalf of each load electricity price on map network node being distributed to its administration respectively by each described load;
Step 10, the position of each load when completing according to described step 9 in the three-dimensional aspect of NETLOGO, and the goal orientation degree of each target of each load, upgrade the initial policy of each load, and according to the method in described step 5, upgrade the set of strategies that each load described is corresponding, then according to the general objective function of each load corresponding, in conjunction with the electricity price that each load is corresponding, obtain the general objective functional value of each strategy in each load its set of strategies corresponding respectively;
Step 11, respectively for each load, judges whether general objective functional value corresponding to the initial policy of load is greater than the general objective functional value in its set of strategies corresponding to other strategy, is then this load stop motion; Otherwise return step 4.
In described step 1, described foundation, based on the union simulation platform of MATLAB and NETLOGO, refers to:
The intelligent grid Multi-Agent simulation platform be made up of MATLAB and NETLOGO, wherein utilizes computing function and the programming technique of MATLAB, sets up the model of power system component and sets up complicated electric power networks simulation model; And NETLOGO is a modeling environment able to programme emulated nature and social phenomenon, be suitable for carrying out modeling to the complication system of Temporal Evolution; Described NETLOGO completes building of power system component general module, and MATLAB carries out every calculating of electric power system, solves the network parameter obtained and realizes information interaction by the interface routine between MATLAB and NETLOGO.
In described step 3, the described general objective function obtaining each flexible load corresponding respectively, its process is:
If economic benefit B kas the income of force device, be defined as follows:
Wherein E kfor the summation of clean input and output, ρ kfor load buys the price of electricity, D kfor load reference power, B kfor economic benefit,
μ kfor the tendency degree of economy, for the tendency degree of comfort level, υ kfor distributed power source sells the price of electricity, G kfor distributed power source reference power;
Definition force device comfort level is as follows:
Wherein C kfor force device comfort level;
The overall utility of force device obtains general objective function representation by two target function weightings, and general objective function definition is as follows:
Wherein R kfor the overall utility of force device.
In described step 4, described is randomly dispersed in each load in the three-dimensional aspect of NETLOGO, and form multiple load bus, and obtain the initial target tendency degree of each target of each load, be the initial policy of each load, its process is:
For the network node in the three-dimensional aspect of described NETLOGO, random setting electricity price, and according to the load bus in the three-dimensional aspect of NETLOGO, set up load agency, the quantity of described load agency is consistent with the quantity of load bus, described load agency and load bus one_to_one corresponding, described each load agency administration each load corresponding, and described each load agency is respectively used to the information transmission between each load and MATLAB of its administration.
In described step 5, described using the initial policy of each load as load datum quantity, respectively for the goal orientation degree of each target of each load, adopt the mode of+i or-i to obtain strategy corresponding to each load respectively, and form the set of strategies of each load in conjunction with the initial policy of each load:
Wherein, i=1, in the three-dimensional aspect of NETLOGO, comprise eight points around each load, these eight points are respectively namely the strategy that each load correspondence eight is different, forms the set of strategies of each load respectively.
The implementation procedure of described step 8 is:
The gross power acted on behalf of by each described load is by the data exchange interface module between MATLAB and NETLOGO, be sent in MATLAB by NETLOGO, the gross power acted on behalf of for each load respectively in MATLAB carries out optimal load flow calculating, obtain the electricity price of generator output and each network node corresponding, and by the electricity price of this each network node, be back in NETLOGO by the data exchange interface module between MATLAB and NETLOGO, upgrade the electricity price on map network node in the three-dimensional aspect of NETLOGO.
In described step 9, described using the electricity price on each network node in three-dimensional for NETLOGO aspect as traction signal, and acted on behalf of each load electricity price on map network node being distributed to its administration respectively by each load, refer to:
Power system dispatching platform often runs once with a set time section, calculate Spot Price when each time period end, predict electricity price, calculating mains frequency and node voltage in short-term, and issue this period electricity price, frequency, voltage to each load agency, large load, issue the history before and after this period and forecasted electricity market price, frequency, voltage when needing simultaneously; Described electricity price, frequency, voltage are referred to as traction signal, instruct self need for electricity of each load adjustment of traction, while maximization number one, serve electrical network.
In described step 10, the described general objective functional value obtaining each strategy in each load its set of strategies corresponding respectively, refers to:
Respectively for each load, judge whether general objective functional value corresponding to the initial policy of load is greater than the general objective functional value in its set of strategies corresponding to other strategy, if be greater than, this load stop motion.
Compared with prior art, the present invention contains following advantage and beneficial effect:
(1) the present invention is according to power system network structure, set up the union simulation platform based on MATLAB and NETLOGO, the data exchange interface module of building between MATLAB and NETLOGO realizes information interaction, propose according to the interactive requirement of net lotus, in conjunction with dissimilar intelligent grid multiple agent characteristic, cooperation control model is set up from the conforming angle of intelligent grid multiple agent multiple-target system, choose appropriate target function when target is various and obtain optimizing operation control mode and parameter, reliability when guaranteeing system cloud gray model and economy, and verify the validity of Optimal Operation Strategies,
(2) the present invention considers that the system containing flexible load multiple agent Collaborative Control has its distinctive feature, and game of vying each other between this kind of model attributes, so adopt intelligent grid distributed consensus optimized algorithm and the Optimal Operation Strategies of the multi-objective coordinated control of multiple agent, on the basis of guaranteeing system reliability, make system have good optimizing operation effect, effectively verify the multi-objective coordinated control and optimize operation result of multiple agent;
(3) the present invention can be widely used in the interactive multi-agent system Controlling model of distributed network lotus, is specially adapted to the intelligent grid multiple agent multiple target consistency optimization method under flexible load.
Accompanying drawing explanation
Fig. 1 is a kind of intelligent grid multiple agent multiple target consistency optimization method flow chart of the present invention.
Fig. 2 is the standardized action space schematic diagram of load k of the present invention.
Fig. 3 is the electrical network Multi-Agent simulation plateform system based on NETLOGO of the present invention.
Embodiment
Below in conjunction with drawings and Examples, the present invention is described in further details.
Figure 1 shows that the flow chart of a kind of intelligent grid multiple agent multiple target consistency optimization method of the present invention.The inventive method is according to the feature of the system of intelligent grid multiple agent Collaborative Control factor, analyze the different characteristic features of different force device, and respective proposed target call, choose appropriate target function when target is various and obtain optimizing operation control mode and parameter, reliability when guaranteeing system cloud gray model and economy, and verify the validity of Optimal Operation Strategies; Implementation step comprises:
Step 1, according to power system network structure, set up the union simulation platform based on MATLAB and NETLOGO, wherein, power system component model is set up in MATLAB, in NETLOGO, definition represents the intelligent body general module of power system component, and meanwhile, the data exchange interface module of building between MATLAB and NETLOGO realizes information interaction.
Described foundation is based on the union simulation platform of MATLAB and NETLOGO, refer to: a kind of intelligent grid Multi-Agent simulation platform be made up of MATLAB and NETLOGO, wherein utilize computing function and the programming technique of MATLAB, set up the model of power system component and set up complicated electric power networks simulation model; And NETLOGO is a modeling environment able to programme emulated nature and social phenomenon, be suitable for carrying out modeling to the complication system of Temporal Evolution; Described NETLOGO completes building of power system component general module, and MATLAB carries out every calculating of electric power system, solves the network parameter obtained and realizes information interaction by the interface routine between MATLAB and NETLOGO.
As shown in Figure 3, power system dispatching platform and NETLOGO many intelligent simulations platform, by MATLAB interface, realize the MAS control of load to the described electrical network Multi-Agent simulation platform be made up of MATLAB and NETLOGO.Power system dispatching platform primary responsibility electricity price calculates and prediction, carries out the Power System Dynamic Simulation of being correlated with simultaneously.For the load emulation based on response electricity price, power system dispatching platform then needs to carry out optimal load flow calculating, obtains the electricity price of now electrical network interdependent node, this electricity price is assigned to the load agency in NETLOGO by NETLOGO and MATLAB interface simultaneously.And NETLOGO emulation platform mainly completes the modeling work of building of operation of power networks environment and electric network element, be embodied in and build topological structure, load agency and load group three layers of operation of power networks environment in NETLOGO; Simultaneously according to the respective characteristic of load group, in NETLOGO, modeling is carried out to its characteristic.NETLOGO and MATLAB interface mainly carries out history electricity price, Spot Price, forecasted electricity market price and the data communication of relevant traction signal between NETLOGO and MATLAB.
Step 2, for various load type, respectively according to load datum quantity, electricity price, and the goal orientation degree of each target of corresponding load, set up the load-Respondence to the Price of Electric Power characteristic model corresponding respectively to various load and power supply type; In electric power system, there is diversified load and power supply, described load comprises rigidity load and flexible load, and described power supply comprises distributed power source and energy-storage travelling wave tube; Wherein, rigidity load refers to the load not participating in electrical network interaction, and flexible load refers to the load participating in electrical network interaction.
Modeling is carried out respectively for load and power supply different qualities.The electricity price perunit value of load power consumption and power supply generating is respectively [ρ k, υ k].In simulation framework proposed by the invention, electricity price is to [ρ k, υ k] be Agent and assign to the traction signal of each load.For the load on different bus, often pair of electricity price may be different.
In the present invention, assuming that the load adopted and power supply are to pursue economy and comfort level for target.In the prior art, the Respondence to the Price of Electric Power characteristic of load and the generating arrangement of power supply do not consider comfort level.Therefore, consider after comfort level, the demand of load cannot according to conventional model Accurate Prediction, and exerting oneself of power supply can not obtain according to traditional generating arrangement method.The behavior that load and power supply pursue respective target can be equivalent to corresponding 2-D space: 1) to the tendency degree μ of economy k, show the maximization avoiding consumer cost on the one hand, obtain on the other hand and maximize economic well-being of workers and staff.2) to the tendency degree of comfort level performance is individual considers self desire and wish, and they understand operative installations or equipment to meet their standard of living (physiology aspect).The goal behavior feature of each load is used describe, A kdifferent values is endowed in economy (as shown in the abscissa of Fig. 2) and comfort level (as shown in the ordinate of Fig. 2) two.
The usefulness that load obtains from electricity consumption can quantize in economic benefit and in comfort level, and the value of these two aspects depends on the individual behavior pattern to power input and output aspect.At this, the power input and output defining described load and power supply are:
1) rigidity load: load q kdo not change with electricity price, namely do not participate in the load of electrical network interaction;
2) flexible load: refer to the load participating in electrical network interaction, wherein d kfor workload demand amount, D kfor load reference power, ρ kfor load buys the price of electricity;
3) distributed power source: wherein g kfor the energy output of distributed power source, G kfor distributed power source reference power, υ kfor distributed power source sells the price of electricity;
4) energy-storage travelling wave tube: during charging be
During electric discharge be
The network power output of load calculates the demand of reference power and comes.In this model, the reference power of load is constant, does not relate to the problem of technical elements.Load-Respondence to the Price of Electric Power characteristic is by the state parameter μ in formula k, determine, wherein said with price ρ k, υ krelevant demand slip and production increase rate are thus cause the Flexible change of power.In addition, contrary with traditional model based on fixing response, in our method, social action, by clear and definite modeling, has considered mutual the brought Flexible change due to social action.
Load space state position as shown in Figure 2.If load k is at position A kplace, represents that this load only considers economic interests: ρ k(υ k) growth can cause minimizing or the increase of power stage.Position B means that it cannot change output or the input variable of electric energy according to the change of price, and it is it is considered that comfort level.Compared with both of these case, not in borderline position, its price is all relevant with comfort level to economic interests to a certain degree.Such as, load k is positioned at a C, says from economy, at price ρ kwhen=1, power demand reduces 0.3, and say from comfort level, power consumption will reduce 0.3.This just means at price ρ kwhen=1, final demand will reduce 0.3* (1-0.3).In like manner at a D place, at price ρ kwhen=1, final demand will reduce 0.7* (1-0.7).At price υ kwhen=1, final energy output will reduce 0.7* (1-0.7).Theoretically, ρ kand υ kcan arrange separately.But for preventing load arbitrage, suppose ρ k=-υ k.
As the income of load, economic benefit B kbe defined as follows:
Wherein E kfor the summation of clean input and output.
Load comfort level is defined as follows:
The overall utility of load obtains general objective function representation by two target function weightings, and general objective function definition is as follows:
Step 3, according to the load-Respondence to the Price of Electric Power characteristic model of the corresponding various load type of the difference set up in described step 2, obtains the target function of each target of each load respectively; And respectively for each load, the target function of each target of load is weighted process, obtains the general objective function of each load corresponding respectively.
The described general objective function obtaining each flexible load corresponding respectively, its process is:
If economic benefit B kas the income of force device, be defined as follows:
Wherein E kfor the summation of clean input and output, ρ kfor load buys the price of electricity, D kfor load reference power, B kfor economic benefit,
μ kfor the tendency degree of economy, for the tendency degree of comfort level, υ kfor distributed power source sells the price of electricity, G kfor distributed power source reference power;
Definition force device comfort level is as follows:
Wherein C kfor force device comfort level;
The overall utility of force device obtains general objective function representation by two target function weightings, and general objective function definition is as follows:
Wherein R kfor the overall utility of force device.
Step 4, is randomly dispersed in each described load in the three-dimensional aspect of NETLOGO, obtains the initial policy of each load; For the network node in the three-dimensional aspect of NETLOGO, set electricity price at random, and set up load agency, its process is:
For the network node in the three-dimensional aspect of described NETLOGO, random setting electricity price, and according to the load bus in the three-dimensional aspect of NETLOGO, set up load agency, the quantity of described load agency is consistent with the quantity of load bus, described load agency and load bus one_to_one corresponding, described each load agency administration each load corresponding, and described each load agency is respectively used to the information transmission between each load and MATLAB of its administration.
Step 5, using the initial policy of each load described as load datum quantity, respectively for the goal orientation degree of each target of each load, adopt the mode of+i or-i to obtain strategy corresponding to each load respectively, and form the set of strategies of each load in conjunction with the initial policy of each load; Described i is each step iteration step length, and described step refers to that electricity price often changes once, and the tactful respective change of load once.
In described step 5, described using the initial policy of each load as load datum quantity, respectively for the goal orientation degree of each target of each load, adopt the mode of+i or-i to obtain strategy corresponding to each load respectively, and form the set of strategies of each load in conjunction with the initial policy of each load:
Wherein, i=1, in the three-dimensional aspect of NETLOGO, comprise eight points around each load, these eight points are respectively namely the strategy that each load correspondence eight is different, forms the set of strategies of each load respectively.
Step 6, adopt the intelligent grid consistency optimized algorithm of the multi-objective coordinated control of multiple agent, respectively coordination computing is optimized to the general objective function of each load, and selects the strategy obtaining each load its maximum general objective functional value corresponding respectively, as the preference policy of each load.
Make x irepresent the state of force device, according to consistency protocol, when the state value of and if only if network opens up all nodes of bowl spares is all equal, the node of this network all reaches unanimously, that is:
x 1＝x 2＝L＝x n
The intelligent grid consistency optimized algorithm of the multi-objective coordinated control of described multiple agent, is adopt distributed Economic Scheduling Policy, refers to:
Under flexible load, the target of Economic Dispatch is that social welfare is maximum.From the angle of distributed optimization, application consistency algorithm, using the incremental cost (IC) of generating set and the increment benefit (IB) of flexible load as consistency variable, Economic Dispatch Problem is solved by the mode of distributed optimization.Be embedded into the local controller in each generating set and flexible load upgrades oneself incremental cost or increment benefit according to the incremental cost of neighbours or increment benefit.One " host groups " and " main load " decision-making is selected whether to increase or reduce overall incremental cost and increment benefit.When the total generated output of generator is greater than load aggregate demand power, will reduce the incremental cost of the overall situation, vice versa.When load aggregate demand power is greater than the total generated output of generator, will increase the increment benefit of the overall situation, vice versa.
This algorithm comprises following process:
Suppose that the cost of electricity-generating function of generating set and the electricity consumption benefit function of flexible load are quadratic function, the cost of electricity-generating function of generating set is as follows:
C i(P Gi)＝α i+β iP Gi+γ iP 2 Gi，i∈S G
The electricity consumption benefit function of flexible load is as follows:
B j(P Dj)＝a j+b jP Dj+c jP 2 Dj，j∈S D
Economic Dispatch Problem refers to that generator and flexible load are under the condition meeting a series of operation constraint, make the optimization problem of the maximization of economic benefit of whole power system operation, that is:
P Gi，min≤P Gi≤P Gi，max，i∈S G
P Dj，min≤P Dj≤P Dj，max，j∈S D
Wherein, P djrepresent the demand power of flexible load j, P girepresent the power output of generating set i.S grepresent generator set, S drepresent flexible load set.Utilize classical method of Lagrange multipliers to solve, make λ represent the Lagrange multiplier corresponding with equality constraint, do not consider constraint, above-mentioned RegionAlgorithm for Equality Constrained Optimization can be converted into:
To variable P gi, P djlocal derviation is asked to obtain optimality condition with λ, that is:
Above formula and the equation of comptability, can obtain according to the equation of comptability:
Namely the optimal solution of economic dispatch makes the incremental cost of generator equal with the increment benefit of flexible load, and wherein m represents generator number, and k represents the number of flexible load.
Suppose that all flexible loads and generating set all run in its power constraints.In this consistency algorithm, the IC of generating set and the IB of flexible load is defined as follows:
Select IC and IB as consistency variable, application consistency algorithm, from the more new formula of the IC of generating set (FollowerGenerator) is:
From the more new formula of the IB of load (FollowerLoad) be:
In order to meet the power-balance constraint in electric power system, represent the difference between flexible load actual demand power and generating set real output with Δ P:
The more new formula of the IC of main generator group (LeaderGenerator) is:
The more new formula of the IB of main load (LeaderLoad) is:
Being wherein convergence coefficient, is a positive scalar.It is relevant with the distributed optimization convergence rate of main generator group and main load.
Step 7, respectively according to the goal orientation degree of each target in the preference policy of each described load, moves to each load respectively in the three-dimensional aspect of NETLOGO on corresponding position, and upgrades the goal orientation degree of each target of each load; Then according to the load-Respondence to the Price of Electric Power characteristic model of correspondence, obtain the power of now each load, and in conjunction with the administration of load agency for corresponding load, obtain the gross power of each load agency respectively.
Step 8, the gross power that each described load is acted on behalf of is sent in MATLAB by NETLOGO, in MATLAB, obtain the electricity price of generator output and each network node corresponding, and be back in NETLOGO, upgrade the electricity price on map network node in the three-dimensional aspect of NETLOGO.Its implementation procedure is:
The gross power acted on behalf of by each described load is by the data exchange interface module between MATLAB and NETLOGO, be sent in MATLAB by NETLOGO, the gross power acted on behalf of for each load respectively in MATLAB carries out optimal load flow calculating, obtain the electricity price of generator output and each network node corresponding, and by the electricity price of this each network node, be back in NETLOGO by the data exchange interface module between MATLAB and NETLOGO, upgrade the electricity price on map network node in the three-dimensional aspect of NETLOGO.
In the present invention, the signal of the interface transmission of MATLAB and NETLOGO is P, f, V and C etc.The C obtained after carrying out electricity price derivation for C, MATLAB is stored in bus matrix, and in three machine nine node systems, electricity price information is the 14th row being stored in Bus matrix, and Bus matrix is as shown in table 1.
Table 1
Connect table
The C of load bus in NETLOGO Calling MATLAB, first will determine the position of this node electricity price at bus matrix, and the Agent of this Nodes obtains the electricity price of this node by interface command statement.On the other hand, Agents also needs by load group gross power by command statement Calling MATLAB, and the existing value of load in electricity price being calculated substitutes and re-starts electricity price calculating.Its code description is as shown in table 2.
Table 2
Step 9, using the electricity price on each network node in three-dimensional for described NETLOGO aspect as traction signal, and is acted on behalf of each load electricity price on map network node being distributed to its administration respectively by each described load.Refer to:
Power system dispatching platform often runs once with a set time section, calculate Spot Price when each time period end, predict electricity price, calculating mains frequency and node voltage in short-term, and issue this period electricity price, frequency, voltage to each load agency, large load, issue the history before and after this period and forecasted electricity market price, frequency, voltage when needing simultaneously; Described electricity price, frequency, voltage are referred to as traction signal, instruct self need for electricity of each load adjustment of traction, while maximization number one, serve electrical network.
In NETLOGO Three Dimensional Interface, initial point is positioned at southwest corner, horizontal direction represents economy tendency degree, vertical direction represents comfort level tendency degree, number range is all 0-1, each user in client layer in the left and right of this aspect, move up and down the change represented respectively economy tendency degree and comfort level tendency degree, while movement, load is also constantly changing, and to the last arrives the some stopping that a general objective is maximum.
Step 10, the position of each load when completing according to described step 9 in the three-dimensional aspect of NETLOGO, and the goal orientation degree of each target of each load, upgrade the initial policy of each load, and according to the method in described step 5, upgrade the set of strategies that each load described is corresponding, then according to the general objective function of each load corresponding, in conjunction with the electricity price that each load is corresponding, obtain the general objective functional value of each strategy in each load its set of strategies corresponding respectively;
The described general objective functional value obtaining each strategy in each load its set of strategies corresponding respectively, refer to: respectively for each load, judge whether general objective functional value corresponding to the initial policy of load is greater than the general objective functional value in its set of strategies corresponding to other strategy, if be greater than, this load stop motion.
Step 11, respectively for each load, judges whether general objective functional value corresponding to the initial policy of load is greater than the general objective functional value in its set of strategies corresponding to other strategy, is then this load stop motion; Otherwise return step 4.
In sum, the system power balance of the interactive electrical network of current multiple agent controls to have become a current study hotspot.The polytropy of the interactive electrical network of multiple agent and uncertainty make it control to become particularly difficulty, can successfully manage based on the conforming method of multiple agent multiple target.In the electrical network MAS control method of design, the module of each force device composition has certain intelligent, external disturbance can be tackled, make active responding, carry out teaching display stand by the communication of self and peripheral modules adjusts with the autonomy acquired a certain degree simultaneously, realize Real-Time Scheduling and distributed scheduling, thus improve reliability and the economy of operation of power networks.Therefore, force device being participated in system power balance, to regard that a multi-agent system studies as be feasible.
The present invention controls in order to the Power Systems balance realizing multiple agent interaction, have employed a kind of electrical network multi-agent modeling, imitation and control scheme.Program emulation platform is made up of NETLOGO and MATLAB, wherein NETLOGO bears the work of Power System Intelligent element model and electrical network MAS control, MATLAB is responsible for every computing of electric power system, realizes whole system network data exchange by the interface module between NETLOGO and MATLAB.In simulating scheme, force device intelligent body and MATLAB transmit interactive information by interface, relevant parameters is uploaded to MATLAB and carries out the every computing of electric power system by force device intelligent body, in Calling MATLAB, fresh signal is assigned to force device intelligent body simultaneously, and each force device intelligent body considers that own target makes active response.
The present invention considers flexible load characteristic, load under different electricity price, to pursue economy and comfort level for target, research distributed consensus Optimized Operation strategy, be optimized coordination computing, systematic function worked in coordination with and reaches optimum effect, the validity of checking Optimal Operation Strategies.By setting up intelligent grid multiple agent multiple-target system cooperation control model, realize indirect, distributed controll.The operation characteristic of accurate description electric network element and system in intelligent grid Complex Networks Theory system, with intelligent body formal description flexible load, the electrical network interaction obtained based on flexible load information interaction runs multiple agent environment, the behavioural norm of the coordination control strategy run based on the electric network element interaction of traction control and the electric network element autonomous operation of Behavior-based control criterion, adopt suitable algorithm and Optimal Operation Strategies, on the basis of guaranteeing system reliability, system is made to have good optimizing operation effect, and verify the multi-objective coordinated control and optimize operation result of multiple agent.

Claims (8)

1. an intelligent grid multiple agent multiple target consistency optimization method, it is characterized in that, according to the feature of the system of intelligent grid multiple agent Collaborative Control factor, analyze the different characteristic features of different force device, and respective proposed target call, choose appropriate target function when target is various and obtain optimizing operation control mode and parameter, reliability when guaranteeing system cloud gray model and economy, and verify the validity of Optimal Operation Strategies; Implementation step comprises:
Step 1, according to power system network structure, set up the union simulation platform based on MATLAB and NETLOGO, wherein, power system component model is set up in MATLAB, in NETLOGO, definition represents the intelligent body general module of power system component, and meanwhile, the data exchange interface module of building between MATLAB and NETLOGO realizes information interaction;
Step 2, for various load type, respectively according to load datum quantity, electricity price, and the goal orientation degree of each target of corresponding load, set up the load-Respondence to the Price of Electric Power characteristic model corresponding respectively to various load and power supply type; Described load comprises rigidity load and flexible load, and described power supply comprises distributed power source and energy-storage travelling wave tube; Wherein, rigidity load refers to the load not participating in electrical network interaction, and flexible load refers to the load participating in electrical network interaction;
Step 3, according to the load-Respondence to the Price of Electric Power characteristic model of the corresponding various load type of the difference set up in described step 2, obtains the target function of each target of each load respectively; And respectively for each load, the target function of each target of load is weighted process, obtains the general objective function of each load corresponding respectively;
Step 4, is randomly dispersed in each described load in the three-dimensional aspect of NETLOGO, obtains the initial policy of each load; For the network node in the three-dimensional aspect of NETLOGO, set electricity price at random, and set up load agency;
Step 5, using the initial policy of each load described as load datum quantity, respectively for the goal orientation degree of each target of each load, adopt the mode of+i or-i to obtain strategy corresponding to each load respectively, and form the set of strategies of each load in conjunction with the initial policy of each load; Described i is each step iteration step length, and described step refers to that electricity price often changes once, and the tactful respective change of load once;
Step 6, adopt the intelligent grid consistency optimized algorithm of the multi-objective coordinated control of multiple agent, respectively coordination computing is optimized to the general objective function of each load, and selects the strategy obtaining each load its maximum general objective functional value corresponding respectively, as the preference policy of each load;
Make x irepresent the state of force device, according to consistency protocol, when the state value of and if only if network opens up all nodes of bowl spares is all equal, the node of this network all reaches unanimously, that is:
x 1＝x 2＝L＝x n
Step 7, respectively according to the goal orientation degree of each target in the preference policy of each described load, moves to each load respectively in the three-dimensional aspect of NETLOGO on corresponding position, and upgrades the goal orientation degree of each target of each load; Then according to the load-Respondence to the Price of Electric Power characteristic model of correspondence, obtain the power of now each load, and in conjunction with the administration of load agency for corresponding load, obtain the gross power of each load agency respectively;
Step 8, the gross power that each described load is acted on behalf of is sent in MATLAB by NETLOGO, in MATLAB, obtain the electricity price of generator output and each network node corresponding, and be back in NETLOGO, upgrade the electricity price on map network node in the three-dimensional aspect of NETLOGO;
Step 9, using the electricity price on each network node in three-dimensional for described NETLOGO aspect as traction signal, and is acted on behalf of each load electricity price on map network node being distributed to its administration respectively by each described load;
Step 10, the position of each load when completing according to described step 9 in the three-dimensional aspect of NETLOGO, and the goal orientation degree of each target of each load, upgrade the initial policy of each load, and according to the method in described step 5, upgrade the set of strategies that each load described is corresponding, then according to the general objective function of each load corresponding, in conjunction with the electricity price that each load is corresponding, obtain the general objective functional value of each strategy in each load its set of strategies corresponding respectively;
Step 11, respectively for each load, judges whether general objective functional value corresponding to the initial policy of load is greater than the general objective functional value in its set of strategies corresponding to other strategy, is then this load stop motion; Otherwise return step 4.
2. a kind of intelligent grid multiple agent multiple target consistency optimization method according to claim 1, it is characterized in that, in described step 1, described foundation, based on the union simulation platform of MATLAB and NETLOGO, refers to:
The intelligent grid Multi-Agent simulation platform be made up of MATLAB and NETLOGO, wherein utilizes computing function and the programming technique of MATLAB, sets up the model of power system component and sets up complicated electric power networks simulation model; And NETLOGO is a modeling environment able to programme emulated nature and social phenomenon, be suitable for carrying out modeling to the complication system of Temporal Evolution; Described NETLOGO completes building of power system component general module, and MATLAB carries out every calculating of electric power system, solves the network parameter obtained and realizes information interaction by the interface routine between MATLAB and NETLOGO.
3. a kind of intelligent grid multiple agent multiple target consistency optimization method according to claim 1, is characterized in that, in described step 3, and the described general objective function obtaining each flexible load corresponding respectively, its process is:
If economic benefit B kas the income of force device, be defined as follows:
Wherein E kfor the summation of clean input and output, ρ kfor load buys the price of electricity, D kfor load reference power, B kfor economic benefit, μ kfor the tendency degree of economy, for the tendency degree of comfort level, υ kfor distributed power source sells the price of electricity, G kfor distributed power source reference power;
Definition force device comfort level is as follows:
Wherein C kfor force device comfort level;
The overall utility of force device obtains general objective function representation by two target function weightings, and general objective function definition is as follows:
Wherein R kfor the overall utility of force device.
4. a kind of intelligent grid multiple agent multiple target consistency optimization method according to claim 1, is characterized in that, in described step 4, described is randomly dispersed in each load in the three-dimensional aspect of NETLOGO, obtains the initial policy of each load; For the network node in the three-dimensional aspect of NETLOGO, set electricity price at random, and set up load agency, its process is:
For the network node in the three-dimensional aspect of described NETLOGO, random setting electricity price, and according to the load bus in the three-dimensional aspect of NETLOGO, set up load agency, the quantity of described load agency is consistent with the quantity of load bus, described load agency and load bus one_to_one corresponding, described each load agency administration each load corresponding, and described each load agency is respectively used to the information transmission between each load and MATLAB of its administration.
5. a kind of intelligent grid multiple agent multiple target consistency optimization method according to claim 1, it is characterized in that, in described step 5, described using the initial policy of each load as load datum quantity, respectively for the goal orientation degree of each target of each load, adopt the mode of+i or-i to obtain strategy corresponding to each load respectively, and form the set of strategies of each load in conjunction with the initial policy of each load:
Wherein, i=1, in the three-dimensional aspect of NETLOGO, comprise eight points around each load, these eight points are respectively namely the strategy that each load correspondence eight is different, forms the set of strategies of each load respectively.
6. a kind of intelligent grid multiple agent multiple target consistency optimization method according to claim 1, it is characterized in that, the implementation procedure of described step 8 is:
The gross power acted on behalf of by each described load is by the data exchange interface module between MATLAB and NETLOGO, be sent in MATLAB by NETLOGO, the gross power acted on behalf of for each load respectively in MATLAB carries out optimal load flow calculating, obtain the electricity price of generator output and each network node corresponding, and by the electricity price of this each network node, be back in NETLOGO by the data exchange interface module between MATLAB and NETLOGO, upgrade the electricity price on map network node in the three-dimensional aspect of NETLOGO.
7. a kind of intelligent grid multiple agent multiple target consistency optimization method according to claim 1, it is characterized in that, in described step 9, described using the electricity price on each network node in three-dimensional for NETLOGO aspect as traction signal, and each load electricity price on map network node being distributed to its administration is acted on behalf of respectively by each load, refer to:
Power system dispatching platform often runs once with a set time section, calculate Spot Price when each time period end, predict electricity price, calculating mains frequency and node voltage in short-term, and issue this period electricity price, frequency, voltage to each load agency, large load, issue the history before and after this period and forecasted electricity market price, frequency, voltage when needing simultaneously; Described electricity price, frequency, voltage are referred to as traction signal, instruct self need for electricity of each load adjustment of traction, while maximization number one, serve electrical network.
8. a kind of intelligent grid multiple agent multiple target consistency optimization method according to claim 1, is characterized in that, in described step 10, the described general objective functional value obtaining each strategy in each load its set of strategies corresponding respectively, refers to:
Respectively for each load, judge whether general objective functional value corresponding to the initial policy of load is greater than the general objective functional value in its set of strategies corresponding to other strategy, if be greater than, this load stop motion.
CN201510759625.0A 2015-11-10 2015-11-10 A kind of intelligent grid multiple agent multiple target uniformity optimization method Active CN105391090B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510759625.0A CN105391090B (en) 2015-11-10 2015-11-10 A kind of intelligent grid multiple agent multiple target uniformity optimization method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510759625.0A CN105391090B (en) 2015-11-10 2015-11-10 A kind of intelligent grid multiple agent multiple target uniformity optimization method

Publications (2)

Publication Number Publication Date
CN105391090A true CN105391090A (en) 2016-03-09
CN105391090B CN105391090B (en) 2017-12-08

Family

ID=55423056

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510759625.0A Active CN105391090B (en) 2015-11-10 2015-11-10 A kind of intelligent grid multiple agent multiple target uniformity optimization method

Country Status (1)

Country Link
CN (1) CN105391090B (en)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106340890A (en) * 2016-09-27 2017-01-18 东南大学 Distributed control method for coordinating charging and discharging efficiency of energy storage systems of power distribution network
CN106886603A (en) * 2017-03-03 2017-06-23 东南大学 The layered distribution type architectural framework and method of a kind of demand response resource optimization
CN108259250A (en) * 2018-02-28 2018-07-06 哈尔滨理工大学 A kind of multiple agent consistency method of sampling
CN108519764A (en) * 2018-04-09 2018-09-11 中国石油大学(华东) Multi-Agent coordination control method based on data-driven
CN108628162A (en) * 2017-03-17 2018-10-09 通用电器技术有限公司 The scalable flexibility control of distributed terminator in power grid
CN114153431A (en) * 2021-11-15 2022-03-08 西安电子科技大学 Large-scale networked software self-optimization device and method based on group intelligence
CN114881489A (en) * 2022-05-13 2022-08-09 重庆邮电大学 Intelligent power grid economic dispatching method based on event triggering and fixed time

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2014128137A (en) * 2012-12-27 2014-07-07 Hitachi Ltd Power system monitoring control device
CN104537178A (en) * 2014-12-31 2015-04-22 南京邮电大学 Electric power system joint simulation modeling method based on Matlab and Netlogo
CN104536304A (en) * 2014-12-31 2015-04-22 南京邮电大学 Electric system load multi-agent control method based on Matlab and Netlogo

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2014128137A (en) * 2012-12-27 2014-07-07 Hitachi Ltd Power system monitoring control device
CN104537178A (en) * 2014-12-31 2015-04-22 南京邮电大学 Electric power system joint simulation modeling method based on Matlab and Netlogo
CN104536304A (en) * 2014-12-31 2015-04-22 南京邮电大学 Electric system load multi-agent control method based on Matlab and Netlogo

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106340890A (en) * 2016-09-27 2017-01-18 东南大学 Distributed control method for coordinating charging and discharging efficiency of energy storage systems of power distribution network
CN106340890B (en) * 2016-09-27 2018-12-28 东南大学 For coordinating the distributed control method of power distribution network energy-storage system efficiency for charge-discharge
CN106886603A (en) * 2017-03-03 2017-06-23 东南大学 The layered distribution type architectural framework and method of a kind of demand response resource optimization
CN106886603B (en) * 2017-03-03 2020-07-14 东南大学 Hierarchical distributed system architecture and method for demand response resource combination optimization
CN108628162A (en) * 2017-03-17 2018-10-09 通用电器技术有限公司 The scalable flexibility control of distributed terminator in power grid
CN108259250A (en) * 2018-02-28 2018-07-06 哈尔滨理工大学 A kind of multiple agent consistency method of sampling
CN108519764A (en) * 2018-04-09 2018-09-11 中国石油大学(华东) Multi-Agent coordination control method based on data-driven
CN114153431A (en) * 2021-11-15 2022-03-08 西安电子科技大学 Large-scale networked software self-optimization device and method based on group intelligence
CN114153431B (en) * 2021-11-15 2024-04-30 西安电子科技大学 Large-scale networked software self-optimization device and method based on group intelligence
CN114881489A (en) * 2022-05-13 2022-08-09 重庆邮电大学 Intelligent power grid economic dispatching method based on event triggering and fixed time

Also Published As

Publication number Publication date
CN105391090B (en) 2017-12-08

Similar Documents

Publication Publication Date Title
CN105391090A (en) Multi-intelligent-agent multi-target consistency optimization method of intelligent power grid
US10607303B2 (en) Coordination of thermostatically controlled loads
CN107579518B (en) Power system environment economic load dispatching method and apparatus based on MHBA
CN111242443B (en) Deep reinforcement learning-based economic dispatching method for virtual power plant in energy internet
CN104536304B (en) A kind of power system load MAS control method based on Matlab and Netlogo
CN107800157A (en) The virtual power plant dual-layer optimization dispatching method of the temperature control load containing polymerization and new energy
Xi et al. A wolf pack hunting strategy based virtual tribes control for automatic generation control of smart grid
Carli et al. A distributed control algorithm for optimal charging of electric vehicle fleets with congestion management
CN106026084B (en) A kind of AGC power dynamic allocation methods based on virtual power generation clan
CN105429185A (en) Economic dispatching method with robust collaborative consistency
MacDougall et al. Mitigation of wind power fluctuations by intelligent response of demand and distributed generation
Golpayegani et al. Multi-agent collaboration for conflict management in residential demand response
US20200409319A1 (en) Multi-period transactive coordination for day-ahead energy and ancillary service market co-optimization with der flexibilities and uncertainties
Hao et al. Reinforcement learning for building energy optimization through controlling of central HVAC system
CN110474353A (en) Layer-stepping energy-storage system and its power grid frequency modulation control method for coordinating of participation
Dong et al. An adaptive decentralized regulation strategy for the cluster with massive inverter air conditionings
Yang et al. Bi-level decentralized control of electric heating loads considering wind power accommodation in real-time electricity market
Dagdougui et al. Towards a concept of cooperating power network for energy management and control of microgrids
CN107392350B (en) Comprehensive optimization method for power distribution network extension planning containing distributed energy and charging stations
CN110991928A (en) Energy management method and system for comprehensive energy system of multiple micro energy networks
CN111242513B (en) Consistency theory-based distributed economic dispatching method for power system
Le Cadre et al. Is energy storage an economic opportunity for the eco-neighborhood?
He et al. Composite socio-technical systems: A method for social energy systems
CN106647249B (en) A kind of power load distributing formula control method based on Netlogo and Matlab
CN114693101B (en) Multi-region thermoelectric coordination control method for multi-agent reinforcement learning and double-layer strategy distribution

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
EE01 Entry into force of recordation of patent licensing contract

Application publication date: 20160309

Assignee: Zhangjiagang Institute of Zhangjiagang

Assignor: NANJING University OF POSTS AND TELECOMMUNICATIONS

Contract record no.: X2020980007094

Denomination of invention: A multi-agent multi-objective consistency optimization method for smart grid

Granted publication date: 20171208

License type: Common License

Record date: 20201026

EE01 Entry into force of recordation of patent licensing contract