CN113362148A - Electric automobile and agent bidding method thereof - Google Patents

Electric automobile and agent bidding method thereof Download PDF

Info

Publication number
CN113362148A
CN113362148A CN202110612631.9A CN202110612631A CN113362148A CN 113362148 A CN113362148 A CN 113362148A CN 202110612631 A CN202110612631 A CN 202110612631A CN 113362148 A CN113362148 A CN 113362148A
Authority
CN
China
Prior art keywords
agent
bidding
electric
user
power
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.)
Pending
Application number
CN202110612631.9A
Other languages
Chinese (zh)
Inventor
方陈
王冰
周健
张秋桥
王皓靖
王敏
时珊珊
刘维扬
刘舒
包海龙
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.)
Hohai University HHU
State Grid Shanghai Electric Power Co Ltd
Original Assignee
Hohai University HHU
State Grid Shanghai Electric Power Co Ltd
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 Hohai University HHU, State Grid Shanghai Electric Power Co Ltd filed Critical Hohai University HHU
Priority to CN202110612631.9A priority Critical patent/CN113362148A/en
Publication of CN113362148A publication Critical patent/CN113362148A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/08—Auctions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • 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
    • G06Q10/06313—Resource planning in a project environment
    • 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
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/0601—Electronic shopping [e-shopping]
    • G06Q30/0611—Request for offers or quotes
    • 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
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/0601—Electronic shopping [e-shopping]
    • G06Q30/0621—Item configuration or customization
    • 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas 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
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Abstract

The invention discloses an electric automobile and a bidding method of an agent thereof, comprising the following steps: s1: respectively clustering user travel time and classifying the initial remaining electric quantity of the electric automobile by using a K mean value clustering algorithm and a quartile method; s2: establishing a bidding interaction model of electric vehicle users and agents, and integrating large-scale charging and discharging resources; s3: establishing a bidding interaction model between an agent and a power dispatching center, wherein the power dispatching center aims at optimizing the load of a power grid and optimizing the benefits of a power grid company, so that the operation cost of the system is reduced; s4: the electric vehicle agent bidding algorithm based on double-layer optimization is provided, the charging and discharging electricity price and the electric quantity obtained by bidding in the electric power market are used as an outer layer optimization model, the benefit balance among all market main bodies is used as an inner layer optimization frame, outer layer optimization is carried out by utilizing a particle swarm optimization, inner layer optimization is carried out by utilizing a genetic algorithm, and the market main bodies mutually achieve optimal bidding. The invention can effectively realize the three-party economic win-win of the user, the agent and the power dispatching center.

Description

Electric automobile and agent bidding method thereof
Technical Field
The invention relates to the field of bidding mechanisms under the characteristic of an electric vehicle cluster, in particular to a research on a bidding interaction model of electric vehicle users and agents and a bidding interaction model of the agents and a power dispatching center, and specifically relates to a bidding method of electric vehicle agents based on double-layer optimization.
Background
In conventional power systems, the power market lacks flexibility. Since the electric power is an irreplaceable commodity, the generator can arbitrarily exercise the right of the generator in the market in market trading, even monopolize the whole electric power market, and earn market violence. Therefore, after the demand side is introduced to participate in market bidding, the power market can be correctly guided to develop in a benign direction. Meanwhile, the trading mechanism provides opportunities for the demand side to participate in market bidding, and the demand side agent can participate in market bidding by integrating the same type of controllable load in the market, so that standby services such as peak regulation, frequency regulation, voltage regulation and the like can be provided for the system, and corresponding compensation benefits are obtained from the standby services. However, the electric power commodity has unique characteristics, and for the purpose of system power balance, safe and reliable operation, a transaction mode of mainly contracting electric power transaction and assisting spot electric power transaction is generally adopted in the transaction type. Generally, the controllable load participates in market bidding to provide auxiliary services for the system, and the proportion of the transaction amount is also determined according to the specific type of the controllable load.
The bidding of the electric automobile agent is one of main ways for the electric automobile to participate in electric power market bidding, and the bidding is suitable for day-ahead trading, real-time trading and auxiliary service trading in market trading according to the driving characteristics and driving requirements of the electric automobile and the time and the trading types of the electric automobile agent participating in the market trading. The bidding transaction of the electric vehicle agent is not limited to the transaction with any electric power market, and not only can be realized in a single market, but also can be realized in a multi-market combined transaction.
The electric vehicle can be regarded as a load when being charged, and can be regarded as a distributed power supply when being idle, and the electric power is transmitted back to the power grid through the electric vehicle access grid (V2G) technology. Proper charge and discharge control can inhibit and eliminate the adverse effect of the electric automobile on the power grid, and can participate in services such as peak regulation, frequency modulation, rotation standby and the like, so that the electric automobile and the power grid develop in a coordinated manner. Therefore, the electric vehicle agent can be used as an intermediary for trading between the user and the power grid, and represent a huge and scattered vehicle owner to intensively schedule the charging and discharging resources of the electric vehicle to participate in bidding in the power market, so that the electric vehicle user, the agent and the power grid can benefit. Therefore, the concept of the power dispatching center is introduced to the power grid side, and the power dispatching center realizes the optimal overall power economic dispatching cost according to the large-scale user operation load integrated by an electric automobile agent, the power purchasing and selling benefits of a power grid company and the unit condition of a generator set.
Disclosure of Invention
The invention aims to provide an electric automobile and an agent bidding method thereof, the method designs a multi-win bidding mechanism by analyzing a bidding mechanism under the characteristic of an electric automobile cluster, establishes a double-layer optimization model for bidding of an electric automobile agent and a user as well as the electric automobile agent and a power dispatching center, and solves the model by adopting a particle swarm and genetic algorithm internal and external optimization relationship to obtain an optimal bidding result, so that the three-win of the user, the electric automobile agent and the power dispatching center can be realized.
In order to achieve the purpose, the technical scheme of the invention is as follows: an electric automobile and an agent bidding method thereof comprise the following steps:
s1: respectively clustering user travel time and classifying the initial residual electric quantity of the electric automobile by using a K-means clustering algorithm and a quartile method to obtain user subgroups with different operating characteristics;
s2: establishing a bidding interaction model of electric vehicle users and agents, taking the time of the electric vehicles accessing and leaving a charging pile as a clustering characteristic, taking the electric vehicles with similar initial electric quantity SOC as another classification standard, and finally taking the charging pile connected with the electric vehicles with similar driving characteristics as a unified user node;
s3: establishing a bidding interaction model between an agent and a power dispatching center, wherein the power dispatching center takes optimization of power grid load and optimization of benefits of a power grid company as a target, and electric vehicle charging and discharging resources dispatched by the power dispatching center are used as a scheme for peak clipping and valley filling;
s4: the electric vehicle agent bidding algorithm based on double-layer optimization is provided, the charging and discharging electricity price and the electric quantity obtained by bidding in the electric power market are used as an outer layer optimization model, the benefit balance among all market main bodies is used as an inner layer optimization frame, outer layer optimization is carried out by utilizing a particle swarm optimization, inner layer optimization is carried out by utilizing a genetic algorithm, and the market main bodies mutually achieve optimal bidding.
The K-means clustering is an unsupervised classification method, the distance is taken as a similarity criterion, and the closer the distance between two objects is, the greater the similarity is.
When N electric automobile samples x exist1,x2,…,xn,…,xNEach sample contains two variables of the travel start time and the travel end time, and N samples need to be divided into K classes, so that the optimization objective of the K-means algorithm is as follows:
Figure BDA0003096529170000031
wherein, mukDenotes the cluster center of the kth class, when xnWhen it belongs to class k, rnk1, otherwise rnkJ denotes the sum of squared errors, and clustering ends when J no longer changes during the iteration of the algorithm.
The K-means algorithm comprises the following specific steps:
step 1.1: randomly generating k clustering centers U ═ U1,u2,…,uk}；
Step 1.2: the euclidean distance of each sample to the cluster center is calculated as follows:
dist(xn,uk)＝||xn,uk||
wherein, if xnTo mukRatio xnIf the distance to other cluster centers is small, the sample x is indicatednBelongs to class k, where r is setnk1, otherwise rnkWhen the error is equal to 0, calculating the sum of squares of the errors J according to a K mean value algorithm;
step 1.3: recalculating cluster center U ═ U1,u2,…,ukIn which μkCan be obtained by the following formula:
Figure BDA0003096529170000032
wherein N isThe total amount of samples of the electric vehicle is new U ═ U1,u2,…,ukAnd (4) repeating the step 1.2 and the step 1.3, finishing clustering when the error square sum J is not changed any more, and at this time, finishing dividing the N electric automobile samples into k types, wherein the travel starting time and the travel finishing time are similar between the samples of each type.
In S2, the distribution of SOC values is determined by quantitive quartile method, and the specific steps are as follows:
step 2.1: recording the vertical time sequence vector of the SOC of the electric automobile as X at the ith sampling pointi＝[xi,1,xi,2,…,xi,n]Wherein i is 1,2, …, n; x is the number ofi,1≤xi,2…≤xi,n-1≤xi,nSecond fraction MiRepresenting SOC longitudinal timing vector XiMedian of (3), MiThe calculation formula is as follows:
Figure BDA0003096529170000041
wherein n is the total sampling amount of the SOC longitudinal time sequence vector of the electric automobile; trisection number represents XiThe numerical values represented by the positions sequentially separating each 25% data point;
step 2.2: the trits divide the SOC vertical timing sequence vectors into 4 classes with equal quantity. When the sampling total n of the SOC longitudinal timing sequence vectors is different, the calculation formulas are respectively as follows:
2.2.1: when n is an even number, the second quantile MiMixing XiIs divided into two subsequences of the same length, denoted by Xi,1＝[xi,1,xi,2,…,xi,(n-1)/2]And Xi,2＝[xi,(n+1)/2,xi,(n+3)/2,…,xn]，Q1,iDenotes the first quantile, Q3,iDenotes the third quantile, Q1,i、Q3,iIs a subsequence Xi,1And Xi,2A median of (d);
2.2.2: when n is 4k +3(k is 0,1,2, …), Q1,i、Q3,iThe calculation formula is as followsShown in the figure:
Figure BDA0003096529170000042
2.2.3: when n is 4k +1(k is 0,1,2, …), Q1,i、Q3,iThe calculation formula is as follows:
Figure BDA0003096529170000043
in S3, an electric vehicle user and agent bidding interaction model is established, wherein the agent objective function is as follows:
Figure BDA0003096529170000051
wherein the content of the first and second substances,
Figure BDA0003096529170000052
the total income of the agent to the Kth user subgroup;
Figure BDA0003096529170000053
the hiring cost given to the agent for the kth subgroup of users;
Figure BDA0003096529170000054
the benefit of discharging to the power grid through the Kth user subgroup is given to the agent;
Figure BDA0003096529170000055
the energy storage operation cost after purchasing electricity for the Kth user subgroup for the agent;
Figure BDA0003096529170000056
compensating the charge for the agent to discharge of the Kth user subgroup; r is the profit sharing proportion given to the user by the agent, and no profit sharing if the agent is lost. In S3, an electric vehicle user and agent bidding interaction model is established, wherein the user targetThe function is as follows:
Figure BDA0003096529170000057
wherein the content of the first and second substances,
Figure BDA0003096529170000058
the total electricity purchasing cost of the Kth user subgroup;
Figure BDA0003096529170000059
giving the agent a profit bonus to the kth subgroup of users;
Figure BDA00030965291700000510
purchasing electricity cost from the agent for the Kth user subgroup;
Figure BDA00030965291700000511
the cost is reduced for the battery of the Kth user subgroup;
Figure BDA00030965291700000512
the K-th subgroup of users is given the hiring cost of the agent.
In S4, a bidding interaction model between the agent and the power dispatching center is established, and the power dispatching center aims to optimize the grid load and optimize the benefits of the grid company, where the power dispatching center aims to optimize the grid load and the grid company, and the grid load variance is the minimum as an objective function according to different charging and discharging requirements of the electric vehicle user for network access, as follows:
Figure BDA00030965291700000513
wherein D isload(k) The power grid load variance is the power grid load variance containing the Kth user subgroup; pbase(t) the initial power grid load value of the electric automobile is not contained in the moment t; pavAnd (t) is a power grid average load value of the electric automobile included in the moment t.
In S4, a bidding interaction model between the agent and the power dispatching center is established, and the power dispatching center aims to optimize the grid load and optimize the benefits of the power grid company, wherein the power dispatching center aims to approximately express the electricity purchasing cost of the power grid company from the electricity generation market as the added cost of the unit combination by dispatching the charging and discharging resources of the electric vehicle, and neglects the participation of models such as the benefit distribution between the unit and the power grid company, and is shown as a function aiming at the lowest electricity economic dispatching cost:
Figure BDA0003096529170000061
Figure BDA0003096529170000062
Figure BDA0003096529170000063
wherein, C'GThe cost is combined for the power system unit without the electric automobile; cGThe cost is combined for a power system unit containing an electric automobile;
Figure BDA0003096529170000064
costs of user discharge amounts purchased by the grid company from the kth subgroup of users through the agents;
Figure BDA0003096529170000065
and the profit of the charging amount of the users sold to the Kth user subgroup by the power grid company through the agent is obtained.
In S5, based on the double-layer optimized electric automobile agent bidding algorithm, in order to realize comprehensive optimization, the invention adopts a linear weighting method in an evaluation function method, integrates the objective functions proposed by S2 and S3 to obtain a single objective function, gives corresponding weight coefficients to each objective according to the importance of each objective, then optimizes the linear combination of the objectives, converts the multi-objective function into the single objective function, and obtains the optimal objective function through the following formula:
Figure BDA0003096529170000066
wherein, O (K) is a single target function of the Kth user subgroup after multi-target normalization processing, and takes the corresponding characteristics of disordered charging as normalization proportion objects to highlight the advantages of the scheduling strategy and disordered charging in the traditional habit mode; the user aims at the lowest cost, contrary to the goal of the maximum profit of the agent, so
Figure BDA0003096529170000067
The sign is taken to be positive,
Figure BDA0003096529170000068
the sign is negative, and the dimensions of the two are the same, both are as follows
Figure BDA0003096529170000069
The normalization is carried out, and the normalization is carried out,
Figure BDA00030965291700000610
the electricity purchasing cost of the k-th user subgroup without the agent during disordered charging; dwx(k) The power grid load difference is generated when the K-th user subgroup is charged in an unordered mode; lambda [ alpha ]1、λ2And λ3The weight coefficients, which are the respective objective functions, represent the relative importance, and can be obtained by the following equation:
Figure BDA00030965291700000611
drawings
FIG. 1 is a block diagram of an optimization algorithm between market entities according to the present invention;
FIG. 2 is a flow chart of the solution of the two-layer optimization model of the present invention;
FIG. 3 is a flow chart of the quartile method calculation employed in the present invention;
FIG. 4 is a clustering result of the travel time of the electric vehicle obtained by the K-means algorithm according to the present invention;
FIG. 5 is a schematic diagram of an initial SOC quartile method according to the present invention using a quartile method;
FIG. 6 is a conventional load graph of the grid of the present invention;
FIG. 7 shows the bidding charging/discharging price results of the electric vehicle agent and the electric network operator according to the present invention;
fig. 8 is a graph showing the results of the optimum charge/discharge amount at each time of the present invention.
Detailed Description
The technical contents, the structural features, the achieved objects and the effects of the present invention are described in detail below with reference to the accompanying drawings.
As shown in fig. 1 and fig. 2, the bidding method for an electric vehicle and its agent provided by the present invention includes the following steps:
s1: respectively clustering user travel time and classifying the initial residual electric quantity of the electric automobile by using a K-means clustering algorithm and a quartile method to obtain user subgroups with different operating characteristics;
s2: establishing a bidding interaction model of electric vehicle users and agents, taking the time of the electric vehicles accessing and leaving a charging pile as a clustering characteristic, taking the electric vehicles with similar initial electric quantity SOC as another classification standard, and finally taking the charging pile connected with the electric vehicles with similar driving characteristics as a unified user node, integrating large-scale charging and discharging resources, and achieving optimized scheduling with the agents;
s3: establishing a bidding interaction model between an agent and a power dispatching center, wherein the power dispatching center aims at optimizing the load of a power grid and optimizing the benefits of a power grid company, and the electric vehicle charging and discharging resources dispatched by the power dispatching center are used as an important scheme for peak clipping and valley filling, so that the system operation cost is reduced, and the optimal whole power economic dispatching cost is realized;
s4: the electric vehicle agent bidding algorithm based on double-layer optimization is provided, the charging and discharging electricity price and the electric quantity obtained by bidding in the electric power market are used as an outer layer optimization model, the benefit balance among all market main bodies is used as an inner layer optimization frame, outer layer optimization is carried out by utilizing a particle swarm optimization, inner layer optimization is carried out by utilizing a genetic algorithm, and the market main bodies mutually achieve optimal bidding.
The single electric vehicle has strong random trip time, but the trip time of large-scale electric vehicles has regularity, so that electric vehicles with similar trip time can be scheduled as one class by using a clustering method in order to realize cluster scheduling of the electric vehicles.
Therefore, in S1, a K-means clustering algorithm is used to perform clustering analysis on the travel start time and the travel end time of the electric vehicles to obtain several electric vehicle subgroups with different travel times, and the electric vehicles in each subgroup have similar travel times, so that the difference of the travel times of the single electric vehicle is eliminated, and a certain electric vehicle subgroup can be studied.
The K-means clustering is an unsupervised classification method, the distance is taken as a similarity criterion, and the closer the two objects are, the greater the similarity is.
Specifically, assume that there are N electric vehicle samples x1,x2,…,xn,…,xNEach sample contains two variables of the travel starting time and the travel ending time, and N samples need to be divided into K classes, so that the optimization target of the K-means algorithm is formula (1):
Figure BDA0003096529170000081
in the formula (1), mukDenotes the cluster center of the kth class, when xnWhen it belongs to class k, rnk1, otherwise rnkJ denotes the sum of squared errors, and clustering ends when J no longer changes during the iteration of the algorithm. The following steps of the algorithm are specifically described:
1) first, k cluster centers U ═ U are randomly generated1,u2,…,uk}；
2) Calculating the Euclidean distance from each sample to the clustering center, wherein the formula (2) is as follows:
dist(xn,uk)＝||xn,uk|| (2)
in the formula (2), if xnTo mukRatio xnIf the distance to other cluster centers is small, the sample x is indicatednBelongs to class k, where r is setnk1, otherwise r nk0. Calculating the sum of squared errors J according to equation (1);
3) recalculating cluster center U ═ U1,u2,…,ukIn which μkCan be obtained by the formula (3):
Figure BDA0003096529170000082
in the formula (3), N is the total number of samples of the electric vehicle, and new U ═ U is obtained1,u2,…,ukAnd after the previous step, repeating the steps 2) and 3), and finishing clustering when the square sum of errors J is not changed any more, wherein at the moment, the N electric automobile samples are divided into k classes, and the travel starting time and the travel finishing time are similar among the samples of each class. Since each sample contains both variables, electric vehicles having similar characteristics are classified according to the characteristics of both variables.
In S2, the distribution of the initial electric quantity (SOC) values of the electric vehicle is determined in a quantitive manner of quartile, and the SOCs are divided into 4 types with equal number, so as to realize statistical classification of the initial charging load SOCs of the electric vehicle.
Recording the vertical time sequence vector of the SOC of the electric automobile as X at the ith sampling pointi＝[xi,1,xi,2,…,xi,n]Wherein i is 1,2, …, n; x is the number ofi,1≤xi,2…≤xi,n-1≤xi,n. Second fraction MiRepresenting SOC longitudinal timing vector XiMedian of (3), MiCan be represented by the formula (4).
Figure BDA0003096529170000091
In the formula (4), n is the total sampling amount of the electric vehicle SOC longitudinal timing sequence vector.
Trisection number represents XiThe three-decimal place number divides the SOC longitudinal timing vector into 4 classes of equal number, represented by the position of each 25% data point. When the total sampling amount n is different, the calculation formulas are respectively as follows:
when n is even number, the second quantile MiMixing XiIs divided into two subsequences of the same length, denoted by Xi,1＝[xi,1,xi,2,…,xi,(n-1)/2]And Xi,2＝[xi,(n+1)/2,xi,(n+3)/2,…,xn]，Q1,iDenotes the first quantile, Q3,iDenotes the third quantile, Q1,i、Q3,iIs a subsequence Xi,1And Xi,2The median of (3).
When n is 4k +3(k is 0,1,2, …), Q is increased1,i、Q3,iCan be obtained by the formula (5):
Figure BDA0003096529170000092
(iii) Q when n is 4k +1(k is 0,1,2, …)1,i、Q3,iCan be obtained by the formula (6):
Figure BDA0003096529170000101
Xiis at a size of XiMiddle 50% of XiAnd (i is a set of 1,2, …, n), the size of the quartile spacing frame reflects the concentration degree of electric vehicle power SOC data as a whole, and the randomness of the electric vehicle travel power consumption is reflected.
From the calculation results of the expressions (4) to (6), X can be obtainediQuartile distance IQR ofiAs shown in formula (7):
IQRi＝Q3,i-Q1,i (7)
in S3, an electric vehicle user and agent bidding interaction model is established, and in order to fully mobilize the enthusiasm of electric vehicle users, a certain proportion red score is introduced between the agent and the user objective function, so that both parties can formulate a flexible benefit distribution scheme while maximizing benefits.
Figure BDA0003096529170000102
The agent objective function is shown as equation (8), where,
Figure BDA0003096529170000103
the total income of the agent to the Kth user subgroup;
Figure BDA0003096529170000104
the hiring cost given to the agent for the kth subgroup of users;
Figure BDA0003096529170000105
the benefit of discharging to the power grid through the Kth user subgroup is given to the agent;
Figure BDA0003096529170000106
the energy storage operation cost after purchasing electricity for the Kth user subgroup for the agent;
Figure BDA0003096529170000107
compensating the charge for the agent to discharge of the Kth user subgroup; r is the profit sharing proportion given to the user by the agent, and no profit sharing if the agent is lost.
Figure BDA0003096529170000108
As shown in equation (9), the user objective function, wherein,
Figure BDA0003096529170000109
the total electricity purchasing cost of the Kth user subgroup;
Figure BDA00030965291700001010
giving the agent the Kth useDividing profit of the family group into red;
Figure BDA00030965291700001011
purchasing electricity cost from the agent for the Kth user subgroup;
Figure BDA00030965291700001012
the cost is reduced for the battery of the Kth user subgroup;
Figure BDA00030965291700001013
the K-th subgroup of users is given the hiring cost of the agent.
In S4, a bidding interaction model between the agent and the power dispatching center is established, and the power dispatching center aims to optimize the load of the power grid and optimize the benefits of the power grid company.
The power dispatching center aims to meet different charging and discharging requirements of electric vehicle users when the electric vehicle users are connected to the network, and the power grid load variance is the minimum as an objective function, as shown in a formula (10):
Figure BDA0003096529170000111
wherein D isload(k) The power grid load variance is the power grid load variance containing the Kth user subgroup; pbase(t) the initial power grid load value of the electric automobile is not contained in the moment t; pavAnd (t) is a power grid average load value of the electric automobile included in the moment t.
The power dispatching center aims to protect the electricity purchasing and selling benefits of power grid companies and agents from being damaged by dispatching charging and discharging resources of the electric automobile, and enables the unit combination cost increased due to the fact that the electric automobile is connected to the network to be the lowest by issuing a power generation load plan, and the optimal overall economic dispatching cost is achieved.
In order to simplify the problem and highlight the transaction at the power distribution side, the invention approximately expresses the electricity purchasing cost of a power grid company from the power generation market as the added cost of the unit combination, neglects the participation of models such as the benefit distribution of the unit and the power grid company and the like, and the formula (11) shows that the function with the lowest power economic dispatching cost as the target is the function:
Figure BDA0003096529170000112
wherein, C'GThe cost is combined for the power system unit without the electric automobile; cGThe cost is combined for a power system unit containing an electric automobile;
Figure BDA0003096529170000113
costs of user discharge amounts purchased by the grid company from the kth subgroup of users through the agents;
Figure BDA0003096529170000114
and the profit of the charging amount of the users sold to the Kth user subgroup by the power grid company through the agent is obtained.
In S5, based on the double-layer optimized electric automobile agent bidding algorithm, in order to realize comprehensive optimization, the invention adopts a linear weighting method in an evaluation function method, integrates the objective functions proposed by S2 and S3 to obtain a single objective function, obtains an optimal solution through the optimization algorithm (namely, endowing the optimal solution with corresponding weight coefficients according to the importance of each objective, then optimizing the linear combination of the objectives, converting a multi-objective function into the single objective function), and obtains the optimal objective function through a formula (12):
Figure BDA0003096529170000115
wherein, O (K) is a single target function of the Kth user subgroup after multi-target normalization processing, and takes the corresponding characteristics of disordered charging as normalization proportion objects to highlight the advantages of the scheduling strategy and disordered charging in the traditional habit mode; the user aims at the lowest cost, contrary to the goal of the maximum profit of the agent, so
Figure BDA0003096529170000121
The sign is taken to be positive,
Figure BDA0003096529170000122
the sign is negative, and the dimensions of the two are the same, both are as follows
Figure BDA0003096529170000123
The normalization is carried out, and the normalization is carried out,
Figure BDA0003096529170000124
the electricity purchasing cost of the k-th user subgroup without the agent during disordered charging; dwx(k) The power grid load difference is generated when the K-th user subgroup is charged in an unordered mode; lambda [ alpha ]1、λ2And λ3The weight coefficients, which are the respective objective functions, represent the relative importance, and can be obtained by equation (13):
Figure BDA0003096529170000125
the solution according to the invention is further illustrated below in a specific example.
Step 1: suppose that a regional electric vehicle agent has 1000 private electric vehicles participating in market bidding, the models of the electric vehicles are shown in table 1, and the number of the vehicles is one third.
TABLE 13 comparison of battery parameters for electric vehicles
Vehicle model Battery capacity/kWh Endurance mileage/km
BYD-E6 63.3 250
Nissan-Leaf 24 160
Long-safety mini 19.2 105
Assuming that the user only charges at home, the user charges at once a day immediately after the travel is finished.
Starting SOC of electric vehicle obeys N (0.3, 0.4)2) Distribution, lowest SOC 10%. The SOC that the user expects to reach is 95%, and the charging power of electric automobile is 7 KW.
According to the K-means clustering algorithm, clustering is carried out according to the starting time and the ending time of the one-day travel of the user, the clustering can be divided into three main classes, the final clustering center is shown in table 2, the scatter diagram of the clustering result is shown in fig. 4, and the number of the main class subgroups respectively accounts for 30%, 20% and 50% of the total number.
TABLE 2 electric vehicle Final Cluster center
Clustering index Main class 1 Main class 2 Main class 3
Start time/h 9.04 14.04 7.60
End time/h 13.32 18.75 19.03
Step 2: according to the quartile method calculation flow chart of fig. 3, the 3 main classes after the travel time is clustered continue to classify the quartile of the electric vehicle SOC.
Based on the coincidence N (0.2, 0.4)2) The initial SOC values (0.1-0.3) of the normally distributed electric quantity are classified by a quartile method, and the quartile point is shown in Table 3.
TABLE 3 initial SOC quartile Point
Quartet site 1 Quartet site 2 Quartile 3
Initial SOC value 0.1608 0.2083 0.2453
As shown in fig. 5, which is a schematic diagram of an initial SOC quartile method obtained by using a quartile method, 3 quantiles are selected from low to high in sequence according to an electric quantity curve and are divided into 4 subclasses according to 25% of the quantity, and the quantities of the subclasses are consistent, so that equal classification of SOC quantities is achieved.
And step 3: parameters in the bidding interaction model of the electric automobile and the agent and the bidding interaction model of the agent and the power dispatching center are set, and the following assumptions are made:
1) the total amount of automobiles in a certain area is 100 thousands, the penetration rate of the electric automobiles is 5%, and all electric automobile users participate in the day-ahead market bidding through the agent V2G. Specific electric vehicle parameters are shown in table 4, where the discharge power is 70% of the charge power.
TABLE 4 electric vehicle-related parameters
Initial SOC Desired SOC Charging and discharging power Capacity of battery
N(0.3,0.42) 0.9-1 7kW and 5kW 36kW·h
2) And taking the trip end and start time of the user as the start and end time of scheduling, and filling the user once a day. 15 minutes is a schedule period and 1 hour is a power rate period. The peak-valley time-of-use electricity prices without electric vehicle bidding are shown in table 5, and the charge price intervals with electric vehicle charge-discharge bidding are shown in table 6.
Further, the value of the discharge price interval is 1.3 times of the charge price interval in consideration of the cost of storage, transportation, breakage and the like of the electric energy.
TABLE 5 electric price table without electric vehicle bidding
Classification Time period Initial price of electricity/(yuan/(kW h))
Peak period 8:00-12:00,17:00-21:00 1.082
Flat time period 12:00-17:00,21:00-24:00 0.687
In the valley period 0:00-8:00 0.365
TABLE 6 optimization interval of charging price of electric vehicle
Time period Peak period In the valley period Flat time period
Upper limit of charging price (yuan/(kW h)) 1.2 0.5 0.8
Lower limit of charging price (yuan/(kW. h)) 0.9 0.3 0.6
3) The unit discharge compensation rate given to the user by the agent is 0.5 yuan/(kW.h); according to the charging and discharging depth of the electric automobile in the embodiment, the per-unit loss cost of the battery is 0.14 yuan/(kW.h); the unit energy storage operation cost of the agent is 0.1 yuan/(kW.h).
4) The profit sharing proportion given to the user by the agent is 0.1; the user gives the agent a hiring fee of 1 yuan/vehicle/day.
5) Because the bidding trading relation among market subjects is based on flexible charging and discharging resources, the optimization control based on the peak clipping and valley filling of the total load is very important, and the weighting coefficients of three objective functions of the user and the agent are lambda1＝λ2＝0.25、λ30.5; the importance degree of the objective function of the user is the same as that of the agent, and the objective importance of the minimum power grid load difference is the maximum.
6) The power grid operator unit consists of 10 thermal power generating units, as shown in table 7, unit output and start-stop time data, as shown in table 8, unit cost coefficient and start-stop expense data, and as shown in fig. 6, the power grid does not contain conventional loads of electric vehicles.
TABLE 7 Unit output, Start and stop time data
Figure BDA0003096529170000141
TABLE 8 Unit cost factor, Start and stop expense data
Figure BDA0003096529170000151
And 4, step 4: as shown in fig. 7, the bidding charging and discharging price of the electric vehicle agent and the electric network operator at each moment is obtained according to the objective function of the bidding algorithm of the electric vehicle agent based on the double-layer optimization. As can be seen from fig. 7, the electric vehicle agent mainly performs charging control in the load and electricity price valley period on the premise of satisfying the driving requirements of each user, and the optimized bid charging price reduces the initial electricity price as a whole, thereby achieving the purpose of reducing the charging cost; the discharge control is mainly carried out during the peak time of the load and the electricity price, and particularly, higher bidding discharge electricity price is achieved during the peak time of the load so as to obtain the optimal discharge income.
The quartile method divides the initial SOC of the electric automobile into 4 classes from low to high, so that the whole electric automobile group is divided into 12 user subgroups. Based on different user travel rules and different initial SOC, the user participates in agent optimized scheduling according to own requirements, and clustering can show that the number of the 3 rd main class accounts for the largest proportion, the travel rule is the most typical, so all 4 subclasses in the 3 rd main class of the user are taken as an example.
As shown in fig. 8, the result graph of the optimal charge/discharge amount at each time point achieved by all 4 subclasses in the user 3 rd main class, that is, the result of the virtual battery charge amount simulation, is shown.
Specifically, a positive value indicates a scheduled charge amount, and a negative value indicates a scheduled discharge amount.
As shown in fig. 8, the charge/discharge scheduling time of the main class user is consistent with the clustering center between about 20 pm and 8 am of the next day; the charging time is mainly in the period from morning to before trip, and the discharging time is concentrated at about 21 moments.
Specifically, the remaining capacity of the class 1 is small, and therefore the charging capacity is large and the discharging capacity is small, and the class 4 is opposite to the class in which the remaining capacity is relatively maximum, and exhibits opposite charging and discharging behaviors.
Finally, a detailed analysis of the economic interests of the market entity is performed.
The electric vehicle user, the agent and the electric power dispatching center mutually achieve the optimal transaction based on respective requirements and benefits, so that the electric power economic dispatching cost and the user electricity purchasing cost are minimum, and the agent business electricity purchasing and selling difference benefit is maximum. As shown in table 9, the optimal benefit for each subject under different scheduling policies.
TABLE 9 optimal benefits for market entities
Figure BDA0003096529170000161
From the economic point of view, although the charge and discharge scheduling control of the agent participating in the user increases the battery loss cost of the user, the overall cost of the user is reduced on the whole by the economic regulation and optimization scheduling such as charge and discharge compensation and profit sharing given to the user, and meanwhile, the agent obtains certain benefits through the price difference of electricity purchase and sale; although the electric vehicle discharge amount purchased by a power grid company in a peak period based on a power dispatching plan increases the electricity purchasing cost, flexible charging and discharging resources provide an additional effective dispatching means for unit combination, the starting and stopping cost is greatly reduced, and the optimal cost of power economic dispatching is realized while the unit combination cost is reduced.
The invention discloses an electric automobile and an agent bidding method thereof, which are based on a K-means algorithm and a quartile method in electric automobile cluster modeling to respectively cluster user travel time and classify initial residual electric quantity of the electric automobile to obtain user subgroups with different operating characteristics, and the overall solution is carried out on a double-layer bidding model by taking a particle swarm algorithm as outer layer optimization and a genetic algorithm as inner layer optimization.
Specifically, each particle represents one charge-discharge time-of-use electrovalence data and one unit combination on-off state; each chromosome represents the residual capacity of each vehicle in a user subgroup.
The inner-layer algorithm firstly optimizes the residual electric quantity of each user subgroup one by one to obtain the optimal charge and discharge control state; then, based on the inner-layer optimizing result, the outer-layer algorithm optimizes the charging and discharging electricity price and the unit combination integrally, so that the market main body mutually achieves the optimal bidding.
In conclusion, the bidding transaction mechanism designed by the project effectively realizes the three-party economic win-win of the user, the agent and the power dispatching center.
While the present invention has been described in detail with reference to the preferred embodiments, it should be understood that the above description should not be taken as limiting the invention. Various modifications and alterations to this invention will become apparent to those skilled in the art upon reading the foregoing description. Accordingly, the scope of the invention should be determined from the following claims.

Claims (10)

1. An electric automobile and an agent bidding method thereof are characterized by comprising the following steps:
s1: respectively clustering user travel time and classifying the initial residual electric quantity of the electric automobile by using a K-means clustering algorithm and a quartile method to obtain user subgroups with different operating characteristics;
s2: establishing a bidding interaction model of electric vehicle users and agents, taking the time of the electric vehicles accessing and leaving a charging pile as a clustering characteristic, taking the electric vehicles with similar initial electric quantity SOC as another classification standard, and finally taking the charging pile connected with the electric vehicles with similar driving characteristics as a unified user node;
s3: establishing a bidding interaction model between an agent and a power dispatching center, wherein the power dispatching center takes optimization of power grid load and optimization of benefits of a power grid company as a target, and electric vehicle charging and discharging resources dispatched by the power dispatching center are used as a scheme for peak clipping and valley filling;
s4: the electric vehicle agent bidding algorithm based on double-layer optimization is provided, the charging and discharging electricity price and the electric quantity obtained by bidding in the electric power market are used as an outer layer optimization model, the benefit balance among all market main bodies is used as an inner layer optimization frame, outer layer optimization is carried out by utilizing a particle swarm optimization, inner layer optimization is carried out by utilizing a genetic algorithm, and the market main bodies mutually achieve optimal bidding.
2. The bidding method for electric vehicles and agents thereof according to claim 1, wherein the K-means clustering is an unsupervised classification method, and the closer the distance between two objects is, the greater the similarity between the two objects is.
3. The method as claimed in claim 2, wherein when there are N electric vehicle samples x, the method comprises1,x2,…,xn,…,xNEach sample contains two variables of the travel start time and the travel end time, and N samples need to be divided into K classes, so that the optimization objective of the K-means algorithm is as follows:
Figure FDA0003096529160000011
wherein, mukDenotes the cluster center of the kth class, when xnWhen it belongs to class k, rnk1, otherwise rnkJ denotes the sum of squared errors, and clustering ends when J no longer changes during the iteration of the algorithm.
4. The bidding method for electric vehicles and agents thereof according to claim 3, wherein the K-means algorithm comprises the following steps:
step 1.1: randomly generating k clustering centers U ═ U1,u2,…,uk}；
Step 1.2: the euclidean distance of each sample to the cluster center is calculated as follows:
dist(xn,uk)＝||xn,uk||
wherein, if xnTo mukRatio xnIf the distance to other cluster centers is small, the sample is indicatedxnBelongs to class k, where r is setnk1, otherwise rnkWhen the error is equal to 0, calculating the sum of squares of the errors J according to a K mean value algorithm;
step 1.3: recalculating cluster center U ═ U1,u2,…,ukIn which μkCan be obtained by the following formula:
Figure FDA0003096529160000021
wherein N is the total number of samples of the electric automobile, and new U ═ U is obtained1,u2,…,ukAnd (4) repeating the step 1.2 and the step 1.3, finishing clustering when the error square sum J is not changed any more, and at this time, finishing dividing the N electric automobile samples into k types, wherein the travel starting time and the travel finishing time are similar between the samples of each type.
5. The bidding method for electric vehicles and their agents as claimed in claim 4, wherein in S2, the distribution of SOC values is determined by quantitive quartile method, which comprises the following steps:
step 2.1: recording the vertical time sequence vector of the SOC of the electric automobile as X at the ith sampling pointi＝[xi,1,xi,2,…,xi,n]Wherein i is 1,2, …, n; x is the number ofi,1≤xi,2…≤xi,n-1≤xi,nSecond fraction MiRepresenting SOC longitudinal timing vector XiMedian of (3), MiThe calculation formula is as follows:
Figure FDA0003096529160000031
wherein n is the total sampling amount of the SOC longitudinal time sequence vector of the electric automobile; trisection number represents XiThe numerical values represented by the positions sequentially separating each 25% data point;
step 2.2: the trits divide the SOC vertical timing sequence vectors into 4 classes with equal quantity. When the sampling total n of the SOC longitudinal timing sequence vectors is different, the calculation formulas are respectively as follows:
2.2.1: when n is an even number, the second quantile MiMixing XiIs divided into two subsequences of the same length, denoted by Xi,1＝[xi,1,xi,2,…,xi,(n-1)/2]And Xi,2＝[xi,(n+1)/2,xi,(n+3)/2,…,xn]，Q1,iDenotes the first quantile, Q3,iDenotes the third quantile, Q1,i、Q3,iIs a subsequence Xi,1And Xi,2A median of (d);
2.2.2: when n is 4k +3(k is 0,1,2, …), Q1,i、Q3,iThe calculation formula is as follows:
Figure FDA0003096529160000032
2.2.3: when n is 4k +1(k is 0,1,2, …), Q1,i、Q3,iThe calculation formula is as follows:
Figure FDA0003096529160000033
6. the electric vehicle and the agent bidding method thereof according to claim 5, wherein in step S3, a bidding interaction model between the electric vehicle user and the agent is established, wherein the agent objective function is as follows:
Figure FDA0003096529160000034
wherein the content of the first and second substances,
Figure FDA0003096529160000035
the total income of the agent to the Kth user subgroup;
Figure FDA0003096529160000036
the hiring cost given to the agent for the kth subgroup of users;
Figure FDA0003096529160000037
the benefit of discharging to the power grid through the Kth user subgroup is given to the agent;
Figure FDA0003096529160000038
the energy storage operation cost after purchasing electricity for the Kth user subgroup for the agent;
Figure FDA0003096529160000039
compensating the charge for the agent to discharge of the Kth user subgroup; r is the profit sharing proportion given to the user by the agent, and no profit sharing if the agent is lost.
7. The bidding method for electric vehicle and its agent as claimed in claim 6, wherein in S3, a bidding interaction model between the user and the agent of the electric vehicle is established, wherein the user objective function is as follows:
Figure FDA0003096529160000041
wherein the content of the first and second substances,
Figure FDA0003096529160000042
the total electricity purchasing cost of the Kth user subgroup;
Figure FDA0003096529160000043
giving the agent a profit bonus to the kth subgroup of users;
Figure FDA0003096529160000044
purchasing electricity cost from the agent for the Kth user subgroup;
Figure FDA0003096529160000045
the cost is reduced for the battery of the Kth user subgroup;
Figure FDA0003096529160000046
the K-th subgroup of users is given the hiring cost of the agent.
8. The electric vehicle and the agent bidding method thereof according to claim 7, wherein in S4, a bidding interaction model between the agent and the power dispatching center is established, and the power dispatching center aims to optimize the grid load and optimize the benefits of the grid company, wherein the power dispatching center aims to obtain a target function with the minimum grid load variance according to different charging and discharging requirements of the electric vehicle users when they enter the network, as follows:
Figure FDA0003096529160000047
wherein D isload(k) The power grid load variance is the power grid load variance containing the Kth user subgroup; pbase(t) the initial power grid load value of the electric automobile is not contained in the moment t; pavAnd (t) is a power grid average load value of the electric automobile included in the moment t.
9. The electric vehicle and the agent bidding method thereof according to claim 8, wherein in S4, a bidding interaction model between the agent and the power dispatching center is established, and the power dispatching center aims to optimize the grid load and optimize the benefits of the grid company, wherein the power dispatching center aims to approximate the electricity purchasing cost of the grid company from the electricity generation market to the added cost of the unit combination by dispatching the charging and discharging resources of the electric vehicle, and neglects the participation of models such as the benefit distribution of the unit and the grid company, and is a function aiming at the lowest electricity economic dispatching cost:
Figure FDA0003096529160000048
Figure FDA0003096529160000049
Figure FDA00030965291600000410
wherein, C'GThe cost is combined for the power system unit without the electric automobile; cGThe cost is combined for a power system unit containing an electric automobile;
Figure FDA0003096529160000051
costs of user discharge amounts purchased by the grid company from the kth subgroup of users through the agents;
Figure FDA0003096529160000052
and the profit of the charging amount of the users sold to the Kth user subgroup by the power grid company through the agent is obtained.
10. The electric vehicle and the agent bidding method thereof according to claim 9, wherein in S5, based on the electric vehicle agent bidding algorithm of the double-layer optimization, in order to achieve the comprehensive optimization, the invention adopts the linear weighting method in the evaluation function method to integrate the objective functions proposed in S2 and S3 to obtain a single objective function, and assigns corresponding weight coefficients to each objective according to the importance of each objective, and then optimizes the linear combination thereof to convert the multi-objective function into the single objective function, and the optimal objective function can be obtained by the following formula:
Figure FDA0003096529160000053
wherein, O (K) is the single eye of the Kth user subgroup after the multi-target normalization processingThe standard function is used for highlighting the advantages of the scheduling strategy and the disordered charging in the traditional habit mode, and the corresponding characteristics of the disordered charging are used as normalization proportion objects; the user aims at the lowest cost, contrary to the goal of the maximum profit of the agent, so
Figure FDA0003096529160000054
The sign is taken to be positive,
Figure FDA0003096529160000055
the sign is negative, and the dimensions of the two are the same, both are as follows
Figure FDA0003096529160000056
The normalization is carried out, and the normalization is carried out,
Figure FDA0003096529160000057
the electricity purchasing cost of the k-th user subgroup without the agent during disordered charging; dwx(k) The power grid load difference is generated when the K-th user subgroup is charged in an unordered mode; lambda [ alpha ]1、λ2And λ3The weight coefficients, which are the respective objective functions, represent the relative importance, and can be obtained by the following equation:
Figure FDA0003096529160000058
CN202110612631.9A 2021-06-02 2021-06-02 Electric automobile and agent bidding method thereof Pending CN113362148A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110612631.9A CN113362148A (en) 2021-06-02 2021-06-02 Electric automobile and agent bidding method thereof

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110612631.9A CN113362148A (en) 2021-06-02 2021-06-02 Electric automobile and agent bidding method thereof

Publications (1)

Publication Number Publication Date
CN113362148A true CN113362148A (en) 2021-09-07

Family

ID=77531109

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110612631.9A Pending CN113362148A (en) 2021-06-02 2021-06-02 Electric automobile and agent bidding method thereof

Country Status (1)

Country Link
CN (1) CN113362148A (en)

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105046354A (en) * 2015-07-09 2015-11-11 国网四川省电力公司经济技术研究院 Multi-agent power distribution network planning scene simulation generation method and system
CN105896674A (en) * 2016-06-01 2016-08-24 中国南方电网有限责任公司电网技术研究中心 Charging control method and system for electric vehicle aggregator
CN106934542A (en) * 2017-03-09 2017-07-07 国网江苏省电力公司电力科学研究院 A kind of electric automobile demand response regulation and control method based on Stark Burger game theory
CN107248010A (en) * 2017-06-06 2017-10-13 重庆大学 The Optimization Scheduling of meter and Load aggregation business and electric automobile response reliability
CN109066663A (en) * 2018-08-31 2018-12-21 国网上海市电力公司 Consider the price competing method of electric car cluster grouping
CN109934487A (en) * 2019-03-11 2019-06-25 国网福建省电力有限公司 A kind of active distribution network coordinated planning method considering multiagent interest game
CN110472785A (en) * 2019-08-08 2019-11-19 西安交通大学 A kind of electric car group's dispatching method based on load classification
CN111619391A (en) * 2020-05-06 2020-09-04 三峡大学 Electric vehicle ordered charging and discharging method based on cooperative game and dynamic time-of-use electricity price
CN111738518A (en) * 2020-06-24 2020-10-02 国家电网公司西南分部 Electric vehicle charging and discharging scheduling method based on average discharge rate
CN112865149A (en) * 2020-12-31 2021-05-28 天津大学 Vehicle network multi-space-time interaction method based on agent and regional time-of-use electricity price

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105046354A (en) * 2015-07-09 2015-11-11 国网四川省电力公司经济技术研究院 Multi-agent power distribution network planning scene simulation generation method and system
CN105896674A (en) * 2016-06-01 2016-08-24 中国南方电网有限责任公司电网技术研究中心 Charging control method and system for electric vehicle aggregator
CN106934542A (en) * 2017-03-09 2017-07-07 国网江苏省电力公司电力科学研究院 A kind of electric automobile demand response regulation and control method based on Stark Burger game theory
CN107248010A (en) * 2017-06-06 2017-10-13 重庆大学 The Optimization Scheduling of meter and Load aggregation business and electric automobile response reliability
CN109066663A (en) * 2018-08-31 2018-12-21 国网上海市电力公司 Consider the price competing method of electric car cluster grouping
CN109934487A (en) * 2019-03-11 2019-06-25 国网福建省电力有限公司 A kind of active distribution network coordinated planning method considering multiagent interest game
CN110472785A (en) * 2019-08-08 2019-11-19 西安交通大学 A kind of electric car group's dispatching method based on load classification
CN111619391A (en) * 2020-05-06 2020-09-04 三峡大学 Electric vehicle ordered charging and discharging method based on cooperative game and dynamic time-of-use electricity price
CN111738518A (en) * 2020-06-24 2020-10-02 国家电网公司西南分部 Electric vehicle charging and discharging scheduling method based on average discharge rate
CN112865149A (en) * 2020-12-31 2021-05-28 天津大学 Vehicle network multi-space-time interaction method based on agent and regional time-of-use electricity price

Similar Documents

Publication Publication Date Title
Zheng et al. Day-ahead bidding strategy for electric vehicle aggregator enabling multiple agent modes in uncertain electricity markets
Hidrue et al. Is there a near-term market for vehicle-to-grid electric vehicles?
Jiang et al. Optimal allocation of shared parking slots considering parking unpunctuality under a platform-based management approach
Zhang et al. EV charging bidding by multi-DQN reinforcement learning in electricity auction market
Aldik et al. A planning model for electric vehicle aggregators providing ancillary services
CN109066663A (en) Consider the price competing method of electric car cluster grouping
CN110852494A (en) Electric vehicle probability load prediction method considering demand response
Jin et al. Optimal aggregation design for massive V2G participation in energy market
Hussain et al. A two-layer decentralized charging approach for residential electric vehicles based on fuzzy data fusion
Zeng et al. Optimal configuration planning of vehicle sharing station-based electro-hydrogen micro-energy systems for transportation decarbonization
CN112766695B (en) Balanced operation method of main and auxiliary combined system under participation of load aggregation main body
Lan et al. Online car-hailing system performance analysis based on Bayesian network
CN113362148A (en) Electric automobile and agent bidding method thereof
CN107154625B (en) Electric car electric discharge electricity price based on fuzzy Bayesian learning negotiates method
Wang et al. Rental strategy for energy storage to participate in frequency regulation market
CN116014773A (en) Scheduling method for two-stage electric vehicle cooperative networking auxiliary service based on deep reinforcement learning
Rodríguez-Molina et al. A model for cost–benefit analysis of privately owned vehicle-to-grid solutions
Parsons et al. Can vehicle-to-grid revenue help electric vehicles on the market
CN114140165A (en) Pricing method of charging and discharging electricity price for considering charging and discharging selection of electric automobile
Zheng et al. The competitiveness measurement of new energy vehicle industry based on grey relational analysis
Ferreira et al. A conceptual V2G aggregation platform
Zhang et al. User Satisfaction and Responsivity Orderly Dispatching Strategies Application on Time-of-use Price of Electronic Vehicle Optimisation
Sutton et al. Strategies to Reduce Congestion and Increase Access to Electric Vehicle Charging Stations at Workplaces
Xu et al. Response capability evaluation model of electric vehicle aggregator based on evolutionary game and response anticipation
Goulias et al. Commercial Fleet Vehicle Additions and Replacements and the Potential Market Penetration for Electric Vehicles

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