CN116054316A

CN116054316A - Ordered charging and discharging method for electric automobile based on chaotic sparrow optimization algorithm

Info

Publication number: CN116054316A
Application number: CN202211520849.2A
Authority: CN
Inventors: 卞海红; 郭正阳; 周辰罡; 张智源; 任权策
Original assignee: Nanjing Institute of Technology
Current assignee: Nanjing Institute of Technology
Priority date: 2022-11-29
Filing date: 2022-11-29
Publication date: 2023-05-02

Abstract

The invention discloses an ordered charging and discharging method of an electric automobile based on a chaotic sparrow optimization algorithm, which models charging loads of three automobile types of private automobiles, taxis and buses in a disordered charging mode; simulating travel time-space distribution of the electric private car by using a Monte Carlo method to obtain schedulable time period data; a multi-objective function is established. Judging whether EV meets the condition of participating in the subsequent ordered charge-discharge optimization, obtaining the prediction results of various indexes, and making a decision whether to respond to V2G; and a simulation experiment is carried out, so that the effectiveness and the rationality of the method are verified. The invention not only ensures the electric quantity requirement of the user side and the economic benefits of discharge, but also takes account of the requirement of the power grid side for reducing peak-valley difference and load variance, and has effectiveness. The method provided by the invention takes a certain response time period of the individual EV as a research object, and the required data can be identified by the equipment end or provided by the user end, so that the method has feasibility and general applicability, and is more practical.

Description

Ordered charging and discharging method for electric automobile based on chaotic sparrow optimization algorithm

Technical Field

The invention belongs to the technical field of electric vehicle charging and discharging, and particularly relates to an electric vehicle ordered charging and discharging method based on a chaotic sparrow optimization algorithm.

Background

The international energy agency published report of the global electric automobile prospect of 2021 indicates that by 2030, the number of electric automobiles on the global road is expected to break through 1.45 hundred million, and consumers in China and Europe are still the most main targets of the world electric automobile market. In recent years, electric vehicles are rushed into the China market on a large scale under the strong promotion of the national power grid company, and China charging infrastructure is rapidly developed. Admittedly, compared with other urban basic loads, the electric vehicle has extremely high charging load, and because the unordered charging of the electric vehicle has extremely strong randomness, the fluctuation of the power grid is increased, and a large peak-valley difference and peak-valley method is generated. Under the background, in order to ensure the travel demands and economic benefits of vehicle owners, the peak-valley difference and variance of a load demand curve are effectively reduced, and the research on the charging and discharging strategies of the electric vehicle shows very high prospective and practicality and has important significance for the protection of ecological environment, the safety of an electric power system and the travel demands of EV users.

Under the multi-time scale, the change of the schedulable capacity of the EV cluster in each time period has strong randomness and fluctuation, most EV users can generate temporary travel demands or prolong parking time due to special conditions, and the actual schedulable capacity in each time period can generate larger deviation compared with a predicted result, so that the accuracy of load demand curve prediction is influenced. At present, most of related researches simulate a load demand curve, and do not sequentially optimize the charge and discharge control coefficients of each response period according to the sequential time sequence of the participation of the EV in the sequential charge and discharge, but in actual situations, the optimization result of the EV which is subsequently connected to a power grid is unknown, which influences the reliability of the calculation result. Similarly, the EV should follow the V2G charge-discharge scheduling on the premise of meeting the basic travel demand, and if the peak shaving demand of the power grid is only used as the main target, the actual experience of the user is ignored, and the reliability of the research result is also affected.

Disclosure of Invention

Aiming at the problems in the prior art, the invention discloses a chaos sparrow optimization algorithm-based ordered charging and discharging method for an electric automobile, comprehensively considers the temporary travel requirements of users, the actual experience of the users after participating in charging and discharging optimization and the peak shaving requirements of a power grid side, and aims to formulate a dedicated charging and discharging optimization scheme aiming at different requirements of EV individual users. In addition, the invention sequentially optimizes the charge and discharge control coefficients of each response period according to the sequential time sequence of EV participation in orderly charge and discharge so as to make the simulation result of the load demand curve have practical significance.

The invention adopts the following technical scheme: the electric automobile ordered charge and discharge method research based on the chaotic sparrow optimization algorithm is characterized by comprising the following steps of: step A, modeling the charge load of three vehicle types, namely private vehicles, taxis and buses, in a disordered charge mode; step B, simulating travel space-time distribution of the electric private car by using a Monte Carlo method based on a regional household vehicle investigation result and a regional traffic network model and combining a Floyd algorithm and a space transfer probability matrix to obtain schedulable time period data; and C, on the basis of peak clipping oriented V2G excitation price, establishing a multi-objective function with the lowest charge and discharge cost of a user, the smallest peak-valley difference of load in a response period, the smallest mean square error of load fluctuation in the response period, the smallest sum of load of each period in the response period after optimization and the average load difference value of day before optimization, and the highest user satisfaction. In addition, five constraints of EV driving mileage constraint, battery charge-discharge capacity constraint, transformer capacity constraint, charge-discharge power constraint and charge-discharge conversion frequency constraint are established; and D, dividing peak-valley time periods according to basic loads in response time periods when the EV is connected to a power grid, centering on charging requirements of users, carrying out charging optimization by correcting a charging and discharging control coefficient in an expected stay time period, and judging whether the EV meets the condition of participating in subsequent ordered charging and discharging optimization. Then, a constraint condition is considered, a chaos sparrow optimization algorithm of the dimension-by-dimension Gaussian variation is adopted to conduct charge and discharge optimization on charge and discharge control coefficients in an EV response period, prediction results of various indexes are obtained, and a user can conveniently make a decision on whether to respond to V2G or not; and E, carrying out simulation experiments on load demands under different charging modes, different optimization weights and different V2G responsivities, and verifying the effectiveness and rationality of the optimization strategy provided by the invention.

Preferably, in the step a, one day is discretized into 96 time periods, and each 15 minutes is a time period. According to traveling purposes, EVs are roughly classified into private cars, taxis and buses.

Preferably, in the step a, the travel chain model may be used to reflect the dynamic characteristics of the travel mode, and the present invention divides the area into five categories, i.e., residential area (H), work area (W), shopping area (S), entertainment area (E), and remaining functional area (O), according to the structure and area function of the travel chain. On the workday, the travel activities of private cars are mainly simple chains (H-W-H), and the travel chain structures containing 3 travel destinations at the maximum are considered, and the structures can be divided into two forms of simple chains and complex chains.

Preferably, in the step a, in the unordered charging mode, a large number of users do not charge the EV every day, so that the charging frequency of the users is taken into consideration, which is beneficial to improving the accuracy of prediction of the Monte Carlo method. It is assumed herein that the frequency of user charging versus daily power consumption is shown in the following table.

TABLE 1

Daily charging frequency	Daily power consumption
		1/7	0％～10％
1/3	10％～20％
		1/2	20％～30％

Preferably, in the step a, in the unordered charging mode on the working day, the private car owner often selects to charge in a parking lot in the residence after going home from work or to perform emergency quick charging near the destination. To reduce battery drain, private car owners often resort to conventional charging during the night. In the ordered charge-discharge mode, most private car owners have low daily electricity, so that abundant electricity is required to respond to urban peak clipping and valley filling. In order to quickly put into operation to earn benefits, taxi drivers often adopt a quick charging mode in a working period, which is also one of main reasons for increasing peak-valley difference of a power grid. In addition, the operation of the electric buses is regular, and a grouping centralized quick charging mode is often adopted. In order to ensure the normal operation of urban traffic, the invention assumes that the bus and the taxi adopt disordered charging modes in the working period. In the chaotic charge mode, the spatiotemporal characteristic parameters of the EV are shown in the following table.

TABLE 2

Preferably, in the step a, based on the habit of charging behavior of the user, a Monte Carlo method is adopted herein to simulate the charge and discharge load of the region EV. Under the situation of disordered charging, the charging load of buses and taxis is very high, and larger impact can be brought to a power grid. Under unordered charging situations, the overall charging requirement of the private car is lower, so that the private car has great schedulable potential, and effective assistance can be provided for peak shaving requirements of the power grid side if the private car is reasonably guided to participate in ordered charging and discharging activities.

Preferably, in the step B, the electric taxis and the electric buses are main sources for causing the peak value of the local time period of the power grid to be too high, but because of actual operation needs, the electric buses adopt a centralized and rapid charging mode, and the electric taxis have a large amount of possibility of participating in ordered charging and discharging activities after the working time period is ended. Considering that private cars have the highest duty ratio in all car types, and most private cars have plentiful time and electric quantity, and meet the basic conditions of participating in ordered charge-discharge scheduling, electric private cars with different capacities are used as research objects, each electric private car is assumed to be equipped with bidirectional V2G equipment, and other electric private cars except the running electric private car can acquire information such as SOC of EV users, willingness and position of participating in V2G peak clipping and valley filling auxiliary service and the like through the communication function of the V2G equipment.

Preferably, in the step B, the distance matrix of the road network is calculated by using Dijkstra algorithm, assuming that Dijkstra shortest path is the total path length of each section of travel chain of the EV.

Dijkstra is mainly characterized by starting from a starting point, adopting a greedy algorithm strategy, traversing adjacent nodes of the top closest to the starting point and not visited each time until the nodes extend to the end point.

Preferably, in the step B, the travel chain is regarded as a markov chain, each travel destination of the EV is regarded as a State, the next destination of the EV is determined by the current State, and the current State is recorded as State _i The next State is State _j ，p _ij To be from State _i Transition to State _j State transition probabilities of (a).

Preferably, in the step B, in order to obtain the V2G schedulable period of each EV, the spatial state distribution of each EV user in each period of time is simulated according to the trip chain, so as to determine whether the EV is in the schedulable state.

Preferably, in the step C, when the price of the discharge service of the EV participating in the peak clipping auxiliary service is formulated, the load condition and the electricity price of the system should be taken into consideration. If EV users participate in peak clipping when the load fluctuation of the power grid is large, the EV users are more profitable. Dividing one day into 96 time periods at 15min as time intervals, and if the EV is connected to the power grid, loading P of the power distribution network in time period t _t Daily average load P greater than power distribution network _average The period is defined as a peak clipping period. In order to embody the peak clipping difference of the EV users participating in the system in different periods, the V2G service compensation price is formulated according to the quality of the EV participating auxiliary service. The invention introduces peak clipping compensation coefficient Q _pen And designing the following compensation coefficient rules:

Q _pen ＝Q _de ·Q _user

Q _user ＝d _j ·d _jb

P _w ＝P _t -P _average

C _pen,t ＝C _p,t ·Q _p

C _pen,t,min ≤C _pen,t ≤C _pen,t,max

in which Q _de For DSO peak clipping demand coefficient, P _w To reduce the system load, Q _user For the user engagement compensation coefficient, mu is the peak clipping demand price compensation coefficient, and the value is 1.1, d _j Adjusting the coefficient, d, for EV user engagement _jb For the engagement of the EV user,

in order to respond to the number of EV users signing up to participate in V2G in time period, M _a For the total number of EV users, C _p,t C is the current city time-of-use electricity price _pen,t Compensating price for discharging service of EV participating in peak clipping auxiliary service, C _pen,t,min 、C _pen,t,max And compensating the price for the minimum and maximum discharging services of the EV participating in the peak clipping auxiliary service respectively.

Preferably, in the step C, the charge-discharge rate, the depth of discharge, the state of charge, the number of cycles, the charge-discharge amount, and the like are main factors affecting the battery loss. The number of battery cycles and the battery life are approximately in a linear relationship, and the battery degradation cost is calculated according to the following formula:

In the method, in the process of the invention,

battery degradation cost (yuan) for the vehicle at time t; a is that _n Taking-0.015625 as a linear relation coefficient between the service life of the battery and the cycle number; c (C) _n Battery capacity for a vehicle; x is x _n,t The method comprises the steps of (1) circularly charging and discharging electric quantity (kW.h) of a vehicle in a t period; c (C) _change Cost (yuan) for battery replacement; s is S _n,t 、S _n,t-1 The SOC of the vehicle in the t period and the last period, respectively. Battery drain costs are incurred when the vehicle responds to V2G discharge in the t period, and no battery drain costs are incurred when the charging operation is performed.

Preferably, in the step C, the EV has 3 states after being connected to the power grid: charge, discharge and silence. In order to conveniently control the charge and discharge behavior of each EV, a charge and discharge control coefficient is introduced:

CG _n ＝[c _n,1 ,c _n,2 ,...,c _n,t ]

wherein: CG (CG) _n The set of EV charge-discharge control coefficients; c _n,t The charge and discharge control coefficient of each period under the EV single parking time length is used for controlling the charge and discharge behavior and the charge and discharge power of the vehicle in each period, and the control rule is as follows:

preferably, in the step C, the objective function is as follows:

(a) Minimum charge and discharge cost of vehicle

Taking the lowest charge and discharge cost of EV in the charge and discharge period as an objective function after the battery cycle charge and discharge cost is considered:

wherein t is _n,in Starting time of EV response charge-discharge scheduling; t is t _n,out The moment at which the EV is expected to leave the grid; g _a To take into account the charge-discharge costs (cells) of the vehicle after battery depletion; p (P) _in The rated charge and discharge power of the vehicle; c (C) _cpeak,t 、C _ca,t 、C _cvalley,t Charging electricity prices (yuan/kW.h) for peak period, flat period, valley period, respectively; c (C) _dpeak,t 、C _da,t 、C _dvalley,t Discharge service prices (yuan/kW.h) in peak period, flat period, valley period, respectively.

(b) Minimum peak-to-valley load difference in response period

In order to avoid the phenomenon of peak-to-peak superposition caused by large-scale charging still performed in the electricity consumption peak period after the EV is connected to the power grid, the aim is to minimize the load peak-to-valley difference in the response period:

ming _b ＝P _n,max -P _n,min

due to the randomness of the trip chain, each EV may respond to multiple V2G schedules during the day. Thus, a study needs to be conducted separately for multiple V2G response periods within a day for each EV. Wherein P is _n,t Charging and discharging power of an nth EV in a t period when responding to a certain V2G schedule; p (P) _n,max 、P _n,max The peak value and the valley value of the electric power used by the nth EV in the response period are respectively.

(c) Minimum mean square error of load fluctuation in response period

The mean square error of the load reflects the fluctuation condition of regional load, and the smaller the mean square error is, the more stable the change trend of the load is indicated. Targeting the minimum mean square error of load fluctuations in the response period:

Wherein t is _n,es The expected duration of the V2G charge-discharge schedule is engaged for the EV.

(d) The sum of the absolute value of the load of each period after the response period is optimized and the average load difference value before the optimization is minimum

In order to reduce the fluctuation of the overall load, the EV peak clipping and valley filling is further guided, and the aim is to minimize the sum of absolute values of the load of each period after the response period is optimized and the average load difference value of the whole day before the optimization:

(e) User satisfaction is highest

The EV user can obtain certain economic benefits in response to the V2G scheduling, but when the EV responds to the V2G, the power grid side needs to ensure normal travel of the user and reduce frequent charge-discharge conversion of the EV in the scheduling process as much as possible to be a main target so as to improve the enthusiasm of the user for responding to the V2G scheduling.

1) User SOC satisfaction lambda

In practical situations, the user may generate temporary travel demands due to emergency, and if the EV is overdischarged in the period of time of the user responding to the V2G, the travel of the user is affected, thereby reducing the enthusiasm of the user responding to the V2G schedule. Thus, a user SOC satisfaction index λ is introduced herein:

wherein S is _n,out SOC at EV departure; s is S _n,in SOC at the start of V2G response for EV; s is S _n,tg Ideal SOC set for the user. As can be seen from the formula, lambda has a value in the range of (0, 1 ]And (3) inner part. The larger the value thereof, the higher the user SOC satisfaction.

2) Degree of satisfaction ρ of charge-discharge exchange times

Frequent battery charging and discharging has a large impact on battery life, which can affect the user's assessment of V2G. Therefore, the charge-discharge exchange number satisfaction index ρ is introduced herein:

wherein n is _n,pe The value range of ρ is (0, 1) as the charge-discharge exchange times of EV in a certain scheduling period]. The larger the value is, the higher the satisfaction degree of the charge and discharge exchange times of the user is.

Based on the above considerations, user satisfaction is defined as the inverse of the product of user SOC satisfaction and charge-discharge exchange number satisfaction: g _e ＝-λρ

From the above equation, the value of EV user satisfaction is within [ -1, 0), and the smaller the value, the higher the user satisfaction.

(f) General objective function

Based on a linear weighting sum method, each objective function is normalized:

wherein g is a multi-objective optimization function;

respectively the maximum value of the single objective function; lambda (lambda) _a 、λ _b 、λ _c 、λ _d 、λ _e The optimization weights of the single objective functions are respectively.

Preferably, in the step C, the constraint conditions are as follows:

(a) EV range constraints

The residual electric quantity of the EV after discharging to the power grid needs to meet the electric quantity requirement of the next stroke, and the discharging quantity is between the maximum value and the minimum value of the battery capacity, namely, the following conditions are met:

Wherein Z is the driving mileage (km) of the next section of the vehicle;

the maximum kilometer energy consumption (kW.h/100 km) of the current vehicle is the maximum.

(b) Battery discharge capacity constraint

The lifetime of an EV battery is related not only to the number of cyclic discharges but also to the depth of discharge, the charge S of the EV in response to the discharge during the discharge period, in order to avoid additional losses to the lifetime of the EV battery _n,t Schedulable power S _n,d The following constraints are made:

wherein t is _n,dstart Is the discharge phase start time; t is t _n,dover Is the discharge phase end time.

(c) Transformer capacity constraint

The overall load of the area is not greater than the upper limit S of the capacity of the area transformer in any time period _T ：

(d) Charge-discharge power constraint

Charge-discharge control coefficient c _n,t The magnitude of (2) affects the charging and discharging power P of EV _n,t And the charging and discharging power is restrained, so that reasonable scheduling on the power grid side is facilitated. Thus, for P _n,t The following constraints are made:

P _n,t ∈{[-10,-5],[5,10],[15,30]}

(e) Constraint of charge-discharge conversion times

Multiple switching of charge and discharge behavior in a short time can negatively impact battery life. Therefore, it is considered that the charge-discharge conversion number n is within the single response period _n,cg Not more than 1/3 of the total length of the time period.

Preferably, in the step D, after the EV is connected to the power grid, the system obtains the battery capacity C of the vehicle through the battery management system of the EV _n Current state of charge S _n,in And records the time t of the vehicle to access the power grid _n,in . In order to facilitate reasonable dispatching of the power grid, a user also needs to input whether to respond to the charge-discharge strategy, and if the user has a response intention, the user also needs to input the expected residence time t of the vehicle _n,stop SOC expected value S at leaving _n,expect . User actual departure time t _n,route And if the user leaving time is greater than the expected leaving time, the system informs the user and requests the user to update the expected leaving time to continue to optimize the charge and discharge. (1) First stage charge optimization centered on user charge demand

Sequencing the simulation results of the schedulable time periods of the electric private car from small to large according to the starting time. If it isAnd responding to the V2G charge-discharge scheduling in a certain schedulable period, wherein the period is the response period. And superposing load prediction results of taxis and buses with the basic load of the power grid, and dividing peak, flat and valley sections on the basis. If the expected residence time t is input by the user _n,stop If the charge state is less than or equal to 3, the charge state is not discharged, if the user has a charge wish, the unordered charge mode is adopted by default until the SOC of the EV reaches the expected state S _n,expect Or the user leaves, otherwise, the control coefficient of the period is corrected to be in a silence state. When the expected stay time t is input by the user _n,stop When the virtual charging time is greater than 3, introducing virtual charging time length

Calculating the user desired state of charge S in constant power charging mode _n,expect Greater than state of charge S when accessing the grid _n,in When the EV electric quantity reaches the user expected state of charge S _n,expect The time required is:

wherein P is _in For charging power, η is charging efficiency. If the expected parking time t input by EV user _n,stop Less than or equal to the virtual charge duration

The system judges that the EV does not meet the ordered charge-discharge condition, and sets a charge-discharge control coefficient in the EV charge period to 1. Otherwise, the system judges that the EV meets the ordered charge-discharge condition, and the EV is arranged to participate in the second-stage ordered charge-discharge optimization.

(2) Chaotic sparrow optimization algorithm based on dimension-by-dimension Gaussian variation and second-stage ordered charge and discharge optimization

When the EV participates in ordered charge and discharge, a chaos sparrow optimization algorithm based on the dimension-by-dimension Gaussian variation is utilized to optimize the charge and discharge control coefficient of the EV. First, a random sparrow is initialized, its dimension d and the expected residence time t _n,stop Similarly, assuming each population consists of n sparrows, initializing population X as:

wherein f is a fitness value. F (F) _x Is a population fitness value.

The population is initialized using singer mapping strategy, the expression of which is as follows:

Wherein x is in the range of [0,1].

After sparrows of the charge and discharge control coefficients are randomly initialized, the charge and discharge control coefficients of the EV in the response period are preset. And considering the battery discharge capacity constraint, the initialized charge and discharge control coefficient needs to be corrected. Introducing virtual SOC variables

Charging and discharging control coefficient c initialized according to EV _n,t Calculating virtual SOC after the charge and discharge behaviors of each period are completed in sequence: />

Wherein S is _n,t-1 Is the actual SOC value of the last period. Virtual SOC variable at a certain time period

When the charge and discharge control coefficient is larger than the threshold value 0.95, if the expected departure time of the user exceeds 3 time periods and the continuous time period does not belong to the valley time period, the charge and discharge control coefficient of the next 3 time periods is corrected to be negative, otherwise, the charge and discharge control coefficient is corrected to be 0. Considering the battery discharge capacity aboutThe SOC satisfaction of the bundle and the user, the virtual SOC variable +.>

When the charge and discharge control coefficient is smaller than the threshold value 0.3, if the expected departure time of the user exceeds 3 time periods, correcting the charge and discharge control coefficient of the next 3 time periods to be positive; if the expected departure time from the user is less than 3 time periods, the charge and discharge control coefficients of the time period and the adjacent time periods are corrected to be positive.

After the correction of the charge and discharge control coefficient is completed, the sparrow population is predated, and the positions of discoverers in the population are updated:

In the formula, t is the iteration number. Iter _max The maximum iteration number is initially set. Q is a random number subject to normal distribution. L is a 1 x d order matrix, and each element in the matrix is 1.R is R ₂ Is an early warning value in the range of 0,1]. ST is a safety value in the range of [0.5,1]. When the early warning value does not exceed the safety value, the discoverer performs a large-range searching action, and when the early warning value exceeds the safety value, all sparrows need to stop the searching action and return to the safety area.

The participants in the sparrow population can get close to and compete for food immediately after the discoverers find food, but can also find food during movement, and when the number of the participants exceeds a certain number, the participants need to go to other places to search for food, and the position update expression is as follows:

wherein x is _p To find the optimal position occupied by the person. X is X _worst Is the current global worst position. A is a 1 x d matrix in which each element has a random amplitude of 1 or-1, and A ⁺ ＝A ^T (AA ^T ) ^-1 。

Since the sparrow-catching person can catch sparrows at the edge of the population, and the sparrows which are aware of danger are set to be 15% to 30% of the total population, the position update formula of the individuals is as follows:

wherein x is _best Is the global optimum. Beta is a step control parameter, and the value of beta is a random number obeying normal distribution of the mean value 0 and the variance 1. K is a random number between-1 and 1. f (f) _i Is the adaptability of sparrow individuals. f (f) _g Is the current global best fitness value. f (f) _w And the current global worst fitness value. Epsilon is a constant, and the occurrence of the case that the denominator is 0 is prevented. After the position of the sparrow is updated, the next iteration is carried out, the position of the sparrow is updated again, all sparrows are updated by utilizing a turning strategy, and the optimal sparrow is subjected to Gaussian variation of the dimension-by-dimension inertia factors:

X _i (t+1)＝x _i (t)+S(r ₁ *x _best -r ₂ *x _i (t))

X _bestnew (j)＝w*X _best (j)+randn*X _best (j)

where S is a null factor, which determines the position of the flap opposite the prey, and takes a value of 2.r is (r) ₁ 、r ₂ Is a random number, and has a value of 0 to 1.w is an inertial weight factor. Repeating the steps until all iterative processes are completed, and finally obtaining the optimal fitness value which is the optimal value of the sparrow population.

And when the user selects to respond to the charging and discharging of the V2G, updating the load demand curve according to the optimal solution result. In general, the actual residence time of the user often has a certain deviation from the expected residence time, and when the user leaves the power grid, the V2G device updates the load demand curve according to the current state of the EV, so as to avoid affecting the accuracy of the ordered charge-discharge optimization of the EV after the EV leaves the EV which is subsequently connected to the power grid.

Preferably, in the step E, the average benefit of ordered charge and discharge is defined as the inverse of the average value of the actual charge and discharge costs of all EV users participating in V2G ordered charge and discharge in each response period in one day. The average SOC satisfaction is the average value of the sum of the SOC satisfaction when the user actually leaves in each response period in one day, the average charge-discharge conversion times are the average value of the sum of the charge-discharge conversion times from the start of the ordered charge-discharge of the response V2G to the actual leave of the user in each response period in one day, and the average optimal fitness value is the average value of the optimal fitness values in each response period in one day.

The beneficial effects are that: the invention has the following beneficial effects:

1. the invention discloses an ordered charging and discharging method of an electric automobile based on a chaotic sparrow optimization algorithm, which considers the temporary travel demands of users under the multi-time scale, can formulate a charging and discharging optimization scheme aiming at different demands of EV individual users, and has high feasibility, wherein the required basic information can be obtained through V2G equipment or users;

2. according to the invention, the charge and discharge control coefficients of each response period are sequentially optimized according to the sequential time sequence of EV participation in ordered charge and discharge, and the simulation result of the load demand curve has more practical significance;

3. the method provided by the invention can bring considerable discharge benefits to EV users on the premise of ensuring travel demands of the EV users, can effectively reduce load variances of load curves, and ensures benefit demands of a power grid side and a user side.

Drawings

FIG. 1 is a flow chart of the present invention;

FIG. 2 is a distribution of various types of electric vehicles;

FIG. 3 is a travel chain basic structure;

FIG. 4 is a simulation result of the charging load of the regional electric vehicle;

FIG. 5 is a schematic diagram of a regional road network;

fig. 6 is a schematic diagram of a travel chain of an electric vehicle coupled to a traffic network;

FIG. 7 is a graph of load demand curve optimization results at 20% responsiveness;

FIG. 8 is a graph of load demand curve optimization results at 40% responsiveness;

FIG. 9 is a graph of load demand curve optimization results at 60% responsiveness;

FIG. 10 is a graph of load demand curve optimization results at 80% responsiveness;

FIG. 11 is a graph of load demand curve optimization results at 100% responsiveness;

FIG. 12 shows the load peak-to-valley optimization rate at different responsivities using the method of the present invention;

FIG. 13 shows the load variance optimization rate at different responsivities using the method of the present invention.

Detailed Description

The invention will be further described with reference to the accompanying drawings.

The invention discloses an electric vehicle ordered charge and discharge method based on a chaotic sparrow optimization algorithm, which is shown in fig. 1 and comprises the following steps:

step A, modeling the charge load of three vehicle types, namely private vehicles, taxis and buses, in a disordered charge mode;

one day was discretized into 96 time periods, one time period every 15 minutes. According to the trip purpose, EVs are roughly classified into private cars, taxis and buses, and the basic information of each type of EV is shown in fig. 1 on the assumption that 300 EVs are total in a certain area. The travel chain model can be used for reflecting the dynamic characteristics of the travel mode, and the travel chain is divided into five types of residential areas (H), working areas (W), shopping areas (S), entertainment areas (E) and other functional areas (O) according to the structure and the regional functions of the travel chain. On weekdays, the travel activities of private cars are mainly simple chains (H-W-H), and the structures of travel chains including 3 travel destinations at maximum are considered to be divided into two forms of simple chains and complex chains, as shown in fig. 2.

In the unordered charging mode, a large number of users do not charge the EV every day, so that the charging frequency of the users is taken into consideration, and the accuracy of Monte Carlo method prediction is improved. It is assumed herein that the frequency of user charging versus daily power consumption is shown in the following table.

TABLE 3 Table 3

In the work day unordered charging mode, private car owners often choose to charge in parking lots in a residential area after going home from work or to charge in emergency and fast near a destination. To reduce battery drain, private car owners often resort to conventional charging during the night. In the ordered charge-discharge mode, most private car owners have low daily electricity, so that abundant electricity is required to respond to urban peak clipping and valley filling. In order to quickly put into operation to earn benefits, taxi drivers often adopt a quick charging mode in a working period, which is also one of main reasons for increasing peak-valley difference of a power grid. In addition, the operation of the electric buses is regular, and a grouping centralized quick charging mode is often adopted. In order to ensure the normal operation of urban traffic, the invention assumes that the bus and the taxi adopt disordered charging modes in the working period. In the chaotic charge mode, the spatiotemporal characteristic parameters of the EV are shown in the following table.

TABLE 4 Table 4

Based on the habit of the user charging behavior, the Monte Carlo method is adopted to simulate the charging and discharging load of the region EV. Under the situation of disordered charging, the charging load of buses and taxis is very high, and larger impact can be brought to a power grid. As shown in fig. 3, four peaks appear in the EV load in one day, and a superposition phenomenon appears in the load peaks of three vehicle types near 2:30 pm, and the total load reaches about 1400kW. Under unordered charging situations, the overall charging requirement of the private car is lower, so that the private car has great schedulable potential, and effective assistance can be provided for peak shaving requirements of the power grid side if the private car is reasonably guided to participate in ordered charging and discharging activities.

Step B, analyzing a 2017 full-beauty household vehicle investigation result and a regional traffic network model based on a Dijkstra algorithm and a space transfer probability matrix, and simulating the time-space distribution of the travel of the electric private car by adopting a Monte Carlo method to obtain schedulable time period data;

the electric taxis and the electric buses are main sources for causing the peak value of the local time period of the power grid to be too high, but because of actual operation needs, the charging requirements of the electric taxis and the electric buses are high, the electric buses adopt a centralized rapid charging mode, and the electric taxis have the possibility of participating in orderly charging and discharging activities in a large number after the working time period is ended. Considering that private cars have the highest duty ratio in all car types, and most private cars have plentiful time and electric quantity, and meet the basic conditions of participating in ordered charge-discharge scheduling, electric private cars with different capacities are used as research objects, each electric private car is assumed to be equipped with bidirectional V2G equipment, and other electric private cars except the running electric private car can acquire information such as SOC of EV users, willingness and position of participating in V2G peak clipping and valley filling auxiliary service and the like through the communication function of the V2G equipment. As shown in fig. 5, the present invention simulates the travel path of EV users in combination with a traffic network model, thereby obtaining a schedulable period of each electric private car.

And assuming that the Dijkstra shortest path is the total path length of each section of travel chain of the EV, calculating a distance matrix of the road network by using a Dijkstra algorithm. Dijkstra is mainly characterized by starting from a starting point, adopting a greedy algorithm strategy, traversing adjacent nodes of the top closest to the starting point and not visited each time until the nodes extend to the end point.

The travel chain is regarded as a Markov chain, each travel destination of the EV is regarded as a State, the next destination of the EV is determined by the current State, and the current State is recorded as State _i The next State is State _j ，p _ij To be from State _i Transition to State _j State transition probabilities of (a).

In order to obtain the V2G schedulable period of each EV, the spatial state distribution of each EV user in each period of time is simulated according to the trip chain, so as to determine whether the EV is in the schedulable state. Daily travel time T of EV user _s Obeying a normal distribution, the probability density function is as follows:

in v _s ＝8.56；

As can be seen from the statistics of the traffic trip survey data of the whole beauty family, the time spent for the entertainment and work type trip activities approximately meets N (2.39,0.73) ² ) And N (5.87,1.2) ² ) Is a normal distribution of (c).

Shopping and other types of travel activities take time to approximately meet the exponential distribution:

For shopping activities, v _s ＝0.56，

For other types of activities, v _s ＝0.45，/>

The specific steps of EV travel space-time distribution simulation based on travel chains are as follows:

(1) Initializing the number M of EVs, wherein m=1, and M represents the serial number of EVs;

(2) For the mth EV, assuming that the initial SOC value of the first trip is a random number uniformly distributed in [0.8,1], the length Q of the trip chain is randomly extracted. The number ratio of simple chains to complex chains is about 2 on a weekday. Let k=1, k is the serial number of a certain section of sub travel chain in the travel chain of EV one day;

(3) According to the current sub travel chain k, a Dijkstra algorithm is used for calculating the travel distance, and the travel duration is estimated approximately according to the route matrix and the average speed of each road section under the corresponding period of the road network;

(4) Assuming that the residential area and the working area are both charged with 10kW normally, the rest of the functional areas are charged with 30kW quickly, and the SOC expected value S of the user _n,expect Is [0.6,1 ]]Random numbers uniformly distributed among the two. Randomly extracting the parking time t according to the travel chain _n,rp User-intended parking duration t _n,stop ；

(5) Judging whether k is equal to Q, if so, turning to step 6, if not, enabling k=k+1, and turning to step 3 according to the destination of the next sub-travel chain randomly according to the transition probability matrix;

(6) Judging whether M is equal to M, if so, ending the process, and if not, making m=m+1, and turning to the step 2.

It is assumed that all functional areas are equipped with bi-directional V2G devices, so that the period in which the EV is expected to stay at a certain place is a schedulable period. The trip chain diagram of the coupling of the EV and the traffic road network using the home of the EV user as the start point and the end point of the daily trip is shown in fig. 6.

And C, on the basis of peak clipping oriented V2G excitation price, establishing a multi-objective function with the lowest charge and discharge cost of a user, the smallest peak-valley difference of load in a response period, the smallest mean square error of load fluctuation in the response period, the smallest sum of load of each period in the response period after optimization and the average load difference value of day before optimization, and the highest user satisfaction. In addition, five constraints of EV driving mileage constraint, battery charge-discharge capacity constraint, transformer capacity constraint, charge-discharge power constraint and charge-discharge conversion frequency constraint are established;

when the price of the discharge service of the EV participating in the peak clipping auxiliary service is formulated, the load condition and the electricity price of the system should be taken into consideration. If EV users participate in peak clipping when the load fluctuation of the power grid is large, the EV users are more profitable. Dividing one day into 96 time periods at 15min as time intervals, and if the EV is connected to the power grid, loading P of the power distribution network in time period t _t Daily average load P greater than power distribution network _average The period is defined as a peak clipping period. In order to embody the peak clipping difference of the EV users participating in the system in different periods, the V2G service compensation price is formulated according to the quality of the EV participating auxiliary service. The invention introduces peak clipping compensation coefficient Q _pen And designing the following compensation coefficient rules:

Q _pen ＝Q _de ·Q _user

Q _user ＝d _j ·d _jb

P _w ＝P _t -P _average

C _pen,t ＝C _c,t ·Q _p

C _pen,t,min ≤C _pen,t ≤C _pen,t,max

The charge and discharge rate, the depth of discharge, the state of charge, the number of cycles, the charge and discharge capacity, and the like are the main factors that affect the battery loss. The number of battery cycles and the battery life are approximately in a linear relationship, and the battery degradation cost is calculated according to the following formula:

in the method, in the process of the invention,

battery degradation cost (yuan) for the vehicle at time t; a is that _n Taking-0.015625 as a linear relation coefficient between the service life of the battery and the cycle number; c (C) _n Battery capacity for a vehicle; x is x _n,t The method comprises the steps of (1) circularly charging and discharging electric quantity (kW.h) of a vehicle in a t period; c (C) _change Cost (yuan) for battery replacement; s is S _n,t 、S _n,t-1 The SOC of the vehicle in the t period and the last period, respectively. When the vehicle is at tThe segment response V2G discharges, which generates battery drain costs, while the charging action does not.

The EV has 3 states after access to the grid: charge, discharge and silence. In order to conveniently control the charge and discharge behavior of each EV, a charge and discharge control coefficient is introduced:

CG _n ＝[c _n,1 ,c _n,2 ,...,c _n,t ]

preferably, in the step C, the objective function is as follows:

(a) Minimum charge and discharge cost of vehicle

(b) Minimum peak-to-valley load difference in response period

ming _b ＝P _n,max -P _n,min

due to the randomness of the trip chain, each EV may respond to multiple V2G schedules during the day. Thus, a study needs to be conducted separately for multiple V2G response periods within a day for each EV. Wherein P is _n,t Charging and discharging power of the mth EV in a t period when responding to a certain V2G scheduling; p (P) _n,max 、P _n,min The peak value and the valley value of the electric power used by the mth EV in the response period are respectively.

(c) Minimum mean square error of load fluctuation in response period

wherein t is _m,es The expected duration of the V2G charge-discharge schedule is engaged for the EV.

(e) User satisfaction is highest

1) User SOC satisfaction lambda

wherein S is _n,out SOC at EV departure; s is S _n,in SOC at the start of V2G response for EV; s is S _n,tg Ideal SOC set for the user. As can be seen from the formula, lambda has a value in the range of (0, 1]And (3) inner part. The larger the value thereof, the higher the user SOC satisfaction.

2) Degree of satisfaction ρ of charge-discharge exchange times

(f) General objective function

Based on a linear weighting sum method, each objective function is normalized:

wherein g is a multi-objective optimization function;

Preferably, in the step C, the constraint conditions are as follows:

(a) EV range constraints

wherein L is the driving mileage (km) of the next section of the vehicle;

(b) Battery discharge capacity constraint

(c) Transformer capacity constraint

(d) Charge-discharge power constraint

P _n,t ∈{[-10,-5],[5,10],[15,30]}

(e) Constraint of charge-discharge conversion times

And D, dividing peak-valley time periods according to basic loads in response time periods when the EV is connected to a power grid, centering on charging requirements of users, carrying out charging optimization by correcting a charging and discharging control coefficient in an expected stay time period, and judging whether the EV meets the condition of participating in subsequent ordered charging and discharging optimization. And then, a constraint condition is considered, a chaos sparrow optimization algorithm of the dimension-by-dimension Gaussian variation is adopted to conduct charge and discharge optimization on the charge and discharge control coefficients in the EV response period, so that the prediction results of all indexes are obtained, and a user can conveniently make a decision on whether to respond to V2G.

When the EV is connected to the power grid, the system acquires the battery capacity C of the vehicle through a battery management system of the EV _n Current state of charge S _n,in And records the time t of the vehicle to access the power grid _n,in . In order to facilitate reasonable dispatching of the power grid, a user also needs to input whether to respond to the charge-discharge strategy, and if the user has a response intention, the user also needs to input the expected residence time t of the vehicle _n,stop SOC expected value S at leaving _n,expect . User actual departure time t _n,route And if the user leaving time is greater than the expected leaving time, the system informs the user and requests the user to update the expected leaving time to continue to optimize the charge and discharge.

(1) First stage charge optimization centered on user charge demand

Sequencing the simulation results of the schedulable time periods of the electric private car from small to large according to the starting time. If the user selects to respond to the V2G charge-discharge schedule within a certain schedulable period, the period is the response period. And superposing load prediction results of taxis and buses with the basic load of the power grid, and dividing peak, flat and valley sections on the basis. If the expected residence time t is input by the user _m,es If the charge state is less than or equal to 3, the charge state is not discharged, if the user has a charge wish, the unordered charge mode is adopted by default until the SOC of the EV reaches the expected state S _n,expect Or the user leaves, otherwise, the control coefficient of the period is corrected to be in a silence state. When the expected stay time t is input by the user _n,stop When the virtual charging time is greater than 3, introducing virtual charging time length

Calculating the user desired state of charge S in constant power charging mode _n,expect Greater than state of charge S when accessing the grid _n,in When the EV electric quantity reaches the user expected state of charge S _n,stop The time required is:

(2) Second-stage ordered charge-discharge optimization of chaotic sparrow optimization algorithm based on dimension-by-dimension Gaussian variation

Wherein f is a fitness value. F (F) _x Is a population fitness value.

wherein x is in the range of [0,1].

After sparrows of the charge and discharge control coefficients are randomly initialized, the charge and discharge control coefficients of the EV in the response period are preset. Taking into account battery discharge capacity constraints, initializingAnd the charge and discharge control coefficient is corrected. Introducing virtual SOC variable S _n v _,t Charging and discharging control coefficient c initialized according to EV _n,t Calculating virtual SOC after the charge and discharge behaviors of each period are completed in sequence:

When the charge and discharge control coefficient is larger than the threshold value 0.95, if the expected departure time of the user exceeds 3 time periods and the continuous time period does not belong to the valley time period, the charge and discharge control coefficient of the next 3 time periods is corrected to be negative, otherwise, the charge and discharge control coefficient is corrected to be 0. Virtual SOC variable +.>

After the correction of the charge and discharge control coefficient is completed, the sparrow population is predated, and the positions of discoverers in the population are updated: ,

in the formula, t is the iteration number. Iter _max The maximum iteration number is initially set. Q is a random number subject to normal distribution. L is a 1 x d order matrix, and each element in the matrix is 1.R is R ₂ Is an early warning value in the range of 0,1]. ST is a safety value in the range of [0.5,1]. When the early warning value does not exceed the safety value, the discoverer performs a large-scale searching action, and when the early warning value exceeds the safety value, all sparrows need to stop the searching actionReturning to the safe area.

X _i (t+1)＝x _i (t)+S(r ₁ *x _best -r ₂ *x _i (t))

X _bestnew (j)＝w*X _best (j)+randn*X _best (j)

And E, carrying out simulation experiments on load demands under different charging modes, different optimization weights and different V2G responsivities, and verifying the effectiveness and rationality of the optimization strategy provided by the invention.

The following is one example of an experiment performed with the method of the present invention:

the software and hardware configuration of the test platform is shown in the following table:

TABLE 5

Assume that the base load within the area complies with the base load distribution of the IEEE33 node power distribution system shown in the following table:

TABLE 6

Time of day	Load (kW)	Time of day	Load (kW)	Time of day	Load (kW)	Time of day	Load (kW)
								01:00	1625.3	07:00	2670.2	13:00	3018.4	19:00	3482.8
02:00	1741.4	08:00	2786.3	14:00	2879.1	20:00	3715.0
								03:00	1973.6	09:00	3018.4	15:00	2786.3	21:00	3018.4
04:00	2205.8	10:00	3250.6	16:00	2438.0	22:00	2554.1
								05:00	2321.9	11:00	3482.8	17:00	2321.9	23:00	2089.7
06:00	2554.1	12:00	3250.6	18:00	3018.4	24:00	1857.5

The number of EVs of each type in the region is assumed to be shown in fig. 1, and the base load in the corresponding period is superimposed with the loads of buses and taxis to update the base load distribution of each period. Battery replacement cost C ^B Taking 1 yuan/(W.h). The total area is 3 transformers with the capacity of 1600 kVA. The charge and discharge efficiency was 0.9. Setting the maximum iteration number of QPSO as 600, the population scale as 50 and the acceleration coefficient z ₁ ＝z ₂ =1.5. The time-sharing electricity prices of the peak, flat and valley periods of the area are established by taking general industrial time-sharing electricity prices of Jiangsu provinces as standards, and are shown in the following table.

TABLE 7

/>

Evaluation criteria: the average benefit of ordered charge and discharge is defined as the inverse of the average of the actual charge and discharge costs of all EV users participating in V2G ordered charge and discharge during each response period during the day. The average SOC satisfaction is the average value of the sum of the SOC satisfaction when the user actually leaves in each response period in one day, the average charge-discharge conversion times are the average value of the sum of the charge-discharge conversion times from the start of the ordered charge-discharge of the response V2G to the actual leave of the user in each response period in one day, and the average optimal fitness value is the average value of the optimal fitness values in each response period in one day.

Experimental results: and analyzing peak regulation effects of the electric private car when the V2G responsivity under different optimization weights is respectively 0%, 20%, 40%, 60%, 80% and 100%. Optimization weights at multi-objective optimization with lambda ₁ :λ ₂ :λ ₃ :λ ₄ :λ ₅ ＝0.2:0.2:0.2:0.2:0.2，λ ₁ :λ ₂ :λ ₃ :λ ₄ :λ ₅ =0.1:0.2:0.3:0.3:0.1 and λ ₁ :λ ₂ :λ ₃ :λ ₄ :λ ₅ For example, =0.1:0.3:0.3:0.2:0.1, the simulation results are shown in fig. 7-11. Under different responsivity, various indexes in the evaluation standards are calculated and obtained as shown in tables 8-12.

TABLE 8 charge-discharge optimization results at 20% responsivity (optimized preload peak-valley difference of 2.4096E+03, optimized preload variance of 3.4516E+07)

TABLE 9 charge-discharge optimization results at 40% responsivity (optimized preload peak-valley difference of 2.3898E+03, optimized preload variance of 3.38776E+07)

Table 10 charge and discharge optimization results at 60% responsivity (optimized preload peak-valley difference of 2.3709E+03, optimized preload variance of 3.211 E+07)

TABLE 11 charge-discharge optimization results at 80% responsivity (optimized preload peak-valley difference of 2.3551E+03, optimized preload variance of 3.269E+07)

TABLE 12 optimization results of charge and discharge at 100% responsivity (optimized preload peak-valley difference of 2.3329E+03, optimized preload variance of 3.1959E+07)

By comparing the graphs, different responsivities can affect the peak shaving effect of V2G and the user experience. In general, when the optimized weights are the same, the higher the V2G responsivity is, the ordered charge and discharge levels are The average benefit and average SOC satisfaction are in an ascending trend, while the average optimal fitness value is in a descending trend. The method is characterized in that under the condition of higher responsiveness, the number of the EVs participating in the V2G is increased, more response time periods of the EVs are overlapped with load peak-valley time periods, and in the load peak time periods, the V2G excitation price is higher, so that the average benefit of ordered charge and discharge of the EVs is improved, and the average optimal fitness value is improved by a certain program. The peak-to-valley difference optimization rate and the load variance optimization rate are shown in fig. 12 and 13 at different responsivities. At responsivities of 20% and 40%, the peak shaving effect of V2G is not apparent. This is because, at lower responsivity, the EV number participating in V2G is smaller, and the response period coinciding with the peak period is also smaller. Meanwhile, the charging and discharging strategy provided by the invention is more prone to ensuring the travel demands of users, and the EV is not overdischarged due to higher peak shaving demands. When the responsivity is more than or equal to 60%, the load peak-valley difference optimization rate and the load variance optimization rate are obviously improved, and when the responsivity is 100%, the load variance optimization rate reaches even 50%, because more charging demands are transferred to valley periods through a charging and discharging optimization strategy in load peak periods. When the responsivity is the same, the peak shaving effect of the V2G and the user experience can show different characteristics due to different optimization weights. In general, when the optimization weight is λ ₁ :λ ₂ :λ ₃ :λ ₄ :λ ₅ When the ratio of the charge and discharge conversion times is lower, the average gain of ordered charge and discharge is higher, and the performance of peak regulation is relatively good. When the optimization weight is

λ ₁ :λ ₂ :λ ₃ :λ ₄ :λ ₅ When=0.1:0.3:0.3:0.2:0.1, the average charge-discharge conversion times are lower, the average gain of ordered charge-discharge is higher, but the performance in peak shaving effect and average SOC satisfaction shows fluctuation. When the optimization weight is

λ ₁ :λ ₂ :λ ₃ :λ ₄ :λ ₅ When =0.2:0.2:0.2:0.2:0.2, the performance on each index exhibits fluctuation. Overall, the higher the responsiveness, the better the performance in terms of various indicators. WhileWhen the responsivity is the same, whether each index under different optimization weights is good or bad does not show regularity.

Under the condition of adopting the charge-discharge optimization strategy provided by the invention, the optimized load demand shows the characteristic of load reduction at peak time and load lifting at valley time. Furthermore, it can be observed that the load demand curves at each responsivity exhibit a strong volatility. On one hand, the strategy provided by the invention takes the charge and discharge control coefficient in the EV response time period as an optimization target, and the basic charge information of the users participating in the V2G has strong randomness, so that the charge and discharge control coefficients in each response time period have larger difference. On the other hand, the actual departure time of the user and the expected departure time often do not coincide in normal cases, and therefore dynamic updating of the grid load demand is relatively frequent. In summary, under the situation that the temporary trip behavior of the user is considered, the method provided by the invention not only ensures the electric quantity requirement and the discharge economic benefit of the user side, but also considers the requirement of reducing peak-valley difference and load variance of the power grid side, and has effectiveness. In addition, the method provided by the invention takes a certain response time period of the individual EV as a research object, and the required data can be identified by the equipment end or provided by the user end, so that the method has feasibility and general applicability, and is more practical.

It is to be understood that the above examples of the present invention are provided by way of illustration only and not by way of limitation of the embodiments of the present invention. Other variations or modifications of the above teachings will be apparent to those of ordinary skill in the art. It is not necessary here nor is it exhaustive of all embodiments. Any modification, equivalent replacement, improvement, etc. which come within the spirit and principles of the invention are desired to be protected by the following claims.

Claims

1. An electric automobile ordered charge and discharge method based on a chaotic sparrow optimization algorithm is characterized by comprising the following steps of:

step B, simulating travel space-time distribution of the electric private car by using a Monte Carlo method based on a regional household vehicle investigation result and a regional traffic network model and combining a Floyd algorithm and a space transfer probability matrix to obtain schedulable time period data;

step C, on the basis of peak clipping oriented V2G excitation price, establishing a multi-objective function with the lowest charge and discharge cost of a user, the minimum load peak-valley difference in a response period, the minimum load fluctuation mean square error in the response period, the minimum sum of the load of each period in the response period after optimization and the average load difference value before optimization, and the highest user satisfaction;

Step D, dividing peak-valley time periods according to basic loads in response time periods when the EV is connected to a power grid, taking a user charging requirement as a center, carrying out charging optimization by correcting a charging and discharging control coefficient in an expected stay time period, judging whether the EV meets the condition of participating in subsequent ordered charging and discharging optimization, and then carrying out charging and discharging optimization on the charging and discharging control coefficient in the EV response time periods by adopting a chaotic sparrow optimization algorithm of a dimension-by-dimension Gaussian variation in consideration of constraint conditions to obtain prediction results of various indexes;

and E, carrying out simulation experiments on load demands under different charging modes, different optimization weights and different V2G responsivities, and verifying the effectiveness and rationality.

2. The method for orderly charging and discharging the electric vehicles based on the chaotic sparrow optimization algorithm according to claim 1, wherein the electric buses adopt a centralized rapid charging mode, each electric private car is provided with bidirectional V2G equipment, and other electric private cars except the running electric private car acquire the SOC of the EV user, the willingness to participate in the V2G peak clipping and valley filling auxiliary service and the position information through the communication function of the V2G equipment.

3. According to claimThe method for orderly charging and discharging the electric automobile based on the chaotic sparrow optimization algorithm as described in claim 1, wherein each driving destination of the EV is regarded as a State, the next destination of the EV is determined by the current State, and the current State is recorded as State _i The next State is State _j ，p _ij To be from State _i Transition to State _j State transition probabilities of (a).

4. The method for orderly charging and discharging of the electric automobile based on the chaotic sparrow optimization algorithm according to claim 1, wherein in order to obtain the V2G schedulable time period of each EV, the spatial state distribution of each EV user in each time period is simulated according to a travel chain, so that whether the EV is in a schedulable state is determined.

5. The method for orderly charging and discharging the electric automobile based on the chaotic sparrow optimization algorithm according to claim 1, wherein in order to avoid the phenomenon of peak-to-peak superposition caused by large-scale charging still performed in a power consumption peak period after the EV is connected to a power grid, the aim is to minimize a load peak-to-valley difference in a response period:

ming _a ＝P _n,max -P _n,min

wherein P is _n,t Charging and discharging power of an nth EV in a t period when responding to a certain V2G schedule; p (P) _n,max 、P _n,min The peak value and the valley value of the electric power used by the nth EV in the response period are respectively.

6. The method for orderly charging and discharging the electric automobile based on the chaotic sparrow optimization algorithm according to claim 1, wherein the load mean square error reflects the fluctuation condition of regional load, and the smaller the mean square error is, the more stable the variation trend of the load is indicated, and the aim of minimum load fluctuation mean square error in a response period is achieved:

7. The method for orderly charging and discharging of the electric automobile based on the chaotic sparrow optimization algorithm according to claim 1, wherein in order to reduce the fluctuation of the overall load, the EV peak clipping and valley filling is further guided, and the aim is to minimize the sum of absolute values of the load in each period after the period optimization and the average load difference value all day before the optimization:

wherein t is _n,in Starting time of EV response charge-discharge scheduling; t is t _n,out For the moment when EV is expected to leave the grid, P _average Is the daily average load of the distribution network.

8. The method for orderly charging and discharging electric vehicles based on chaotic sparrow optimization algorithm according to claim 1, wherein the electric quantity S of the EV in response to the discharge in the discharge period is _n,t Schedulable power S _n,d The following constraints are made:

9. According toThe method for orderly charging and discharging of an electric vehicle based on a chaotic sparrow optimization algorithm as claimed in claim 1, wherein after the EV is connected to a power grid, the system obtains the battery capacity C of the vehicle through a battery management system of the EV _n Current state of charge S _n,in And records the time t of the vehicle to access the power grid _n,in The user also needs to input whether to respond to the charge-discharge strategy, and if the user has a response intention, the user also needs to input the expected residence time t of the vehicle _n,stop SOC expected value S at leaving _n,expect User actual departure time t _n,route Having randomness, if the user leaving time is greater than the expected leaving time, the system informs the user and requests the user to update the expected leaving time to continue the charge and discharge optimization,

first stage charge optimization centered on user charge demand:

sequencing the simulation results of the schedulable time periods of the electric private car from small to large according to the starting time, if the user selects to respond to the V2G charge-discharge scheduling in a certain schedulable time period, the time period is the response time period, the load prediction results of the taxis and the buses are overlapped with the power grid base load, the peak, flat and valley sections are divided on the basis, and if the expected residence time t is input by the user _n,stop If the charge state is less than or equal to 3, the charge state is not discharged, if the user has a charge wish, the unordered charge mode is adopted by default until the SOC of the EV reaches the expected state S _n,expect Or the user leaves, otherwise, the control coefficient of the period is revised to be in a silence state,

when the expected stay time t is input by the user _n,stop When the virtual charging time is greater than 3, introducing virtual charging time length

wherein P is _in For the charging power, η is the charging efficiency, if the expected parking time t is input by the EV user _n,stop Less than or equal to the virtual charge duration

The system judges that the EV does not meet the ordered charge-discharge condition, the charge-discharge control coefficient in the EV charge period is set to be 1, otherwise, the system judges that the EV meets the ordered charge-discharge condition, and the EV is arranged to participate in the second-stage ordered charge-discharge optimization;

chaotic sparrow optimization algorithm based on dimension-by-dimension Gaussian variation, and second-stage ordered charge and discharge optimization:

when the EV participates in ordered charge and discharge, a chaos sparrow optimization algorithm based on a dimension-by-dimension Gaussian variation is utilized to optimize the charge and discharge control coefficient of the EV, firstly, a random sparrow is initialized, and the dimension d and the expected residence time t of the random sparrow are calculated _n,stop Similarly, assuming each population consists of n sparrows, initializing population X as:

wherein F is a fitness value, F _x For the fitness value of the population,

Wherein x is in the range of 0,1,

sparrow channel with charge and discharge control coefficientAfter random initialization, the charge and discharge control coefficient of EV in the response period is preset, the charge and discharge control coefficient after initialization is required to be corrected by considering the battery discharge capacity constraint, and virtual SOC variable is introduced

Charging and discharging control coefficient c initialized according to EV _n,t Calculating virtual SOC after the charge and discharge behaviors of each period are completed in sequence:

wherein S is _n,t-1 As the virtual SOC variable of a certain period, the virtual SOC variable is the actual SOC value of the previous period

When the preset time is greater than the threshold value 0.95, if the expected departure time of the user exceeds 3 time periods and the continuous time period does not belong to the valley time period, the charge and discharge control coefficient of the next 3 time periods is corrected to be negative, otherwise, the charge and discharge control coefficient is corrected to be 0, and when the virtual SOC variable +_of a certain time period is considered in consideration of the battery discharge capacity constraint and the SOC satisfaction of the user>

When the charge and discharge control coefficient is smaller than the threshold value 0.3, if the expected departure time of the user exceeds 3 time periods, correcting the charge and discharge control coefficient of the next 3 time periods to be positive; if the expected departure time from the user is less than 3 time periods, the charge and discharge control coefficients of the time period and the adjacent time periods are corrected to be positive,

Wherein t is the iteration number, a.iter _max For initial purposesThe set maximum iteration number, Q is a random number obeying normal distribution, L is a 1 x d-order matrix, each element in the matrix is 1, R ₂ Is an early warning value in the range of 0,1]ST is a safety value in the range of [0.5,1]When the early warning value does not exceed the safety value, the discoverer performs a large-scale searching action, when the early warning value exceeds the safety value, all sparrows need to stop searching actions to return to the safety area,

wherein x is _p To find the optimal position occupied by the person, X _worst For the current global worst position, A is a 1×d order matrix, where each element has a random magnitude of 1 or-1, and A ⁺ ＝A ^T (AA ^T ) ^-1 ，

wherein x is _best For the global optimum position, beta is the step control parameter, the value is the random number obeying the normal distribution of the mean value 0 and the variance 1, K is the random number between-1 and 1, f _i Is the fitness of sparrow individuals, f _g For the current global best fitness value, f _w For the current global worst fitness value, epsilon is a constant, the situation that the denominator is 0 is prevented from occurring, the next iteration is carried out after the position of the sparrow is updated, the position of the sparrow is updated again, and all the sparrows are turned overUpdating the fighting strategy, and carrying out Gaussian variation of the dimensionality-by-dimensionality inertia factors on the optimal sparrows:

X _i (t+1)＝x _i (t)+S(r ₁ *x _best -r ₂ *x _i (t))

X _bestnew (j)＝w*X _best (j)+randn*X _best (j)

wherein S is a null factor, determining the position of turning to the opposite side of the prey, and taking the value of 2, r ₁ 、r ₂ And (3) taking the random number as a random number, wherein the value is 0 to 1, w is an inertia weight factor, and repeating the steps until all iteration processes are completed, and finally obtaining the optimal fitness value which is the optimal value of the sparrow population.

10. The method for orderly charging and discharging of the electric vehicle based on the chaotic sparrow optimization algorithm according to claim 1, wherein average gain of orderly charging and discharging is defined as the opposite number of the average value of the actual charging and discharging costs of all EV users participating in V2G orderly charging and discharging in each response period in one day, average SOC satisfaction is the average value of the sum of the SOC satisfaction when the users actually leave in each response period in one day, average charging and discharging conversion times is the average value of the sum of the charging and discharging conversion times from starting to orderly charging and discharging of the response V2G in each response period in one day to the actual leaving of the users, and average optimal fitness value is the average value of the optimal fitness values in each response period in one day.