CN112347615A

CN112347615A - Power distribution network hybrid optimization scheduling method considering light storage and fast charging integrated station

Info

Publication number: CN112347615A
Application number: CN202011125344.7A
Authority: CN
Inventors: 刘一欣; 吴国诚; 郭力; 胡代豪; 倪筹帷
Original assignee: Tianjin University; Electric Power Research Institute of State Grid Zhejiang Electric Power Co Ltd
Current assignee: Tianjin University; Electric Power Research Institute of State Grid Zhejiang Electric Power Co Ltd
Priority date: 2020-10-20
Filing date: 2020-10-20
Publication date: 2021-02-09

Abstract

The invention discloses a power distribution network hybrid optimization scheduling method considering a light storage and fast charging integrated station, which comprises the steps of establishing a road traffic network model, a driving speed and path model of an electric automobile and an electric automobile travel time model, and establishing a charging load model of the fast charging integrated station based on the three models; based on the charging load model, the uncertain set is adopted to represent the range of photovoltaic output and original load in the power distribution network, the minimum overall operation cost of the power distribution network to which the rapid charging integrated station belongs is taken as a target function, and constraint conditions are set to establish a robust optimization scheduling model of the power distribution network of the rapid charging integrated station; decomposing the model into a main problem model and a sub problem model; the worst scene of the power distribution network tide operation is obtained by optimizing the photovoltaic output and the total time interval of the boundary value of the original load selection interval in the scheduling period; and solving the optimal solution of the main problem model and the sub problem model by adopting a column constraint generation algorithm. The method can flexibly realize the decision of the energy storage system of the quick charging station under the uncertain environment.

Description

Power distribution network hybrid optimization scheduling method considering light storage and fast charging integrated station

Technical Field

The invention relates to a power distribution network scheduling method of an optical storage and fast charging integrated station, in particular to a power distribution network hybrid optimization scheduling method considering the optical storage and fast charging integrated station.

Background

At present, Electric Vehicles (EVs) as a new generation of green Vehicles have great advantages and potentials in the aspects of reducing automobile exhaust emission, relieving energy crisis and the like. The permeability of the charging load in the power distribution network is greatly improved, and the uncertainty of the charging load, particularly the intermittency and randomness of the quick charging load, causes the reliable and economic operation of the power distribution system to face certain risks. In order to reduce the impact of the fast charging load on the power distribution system as much as possible, how to carry out detailed modeling on the fast charging load and an optimal scheduling method for accessing the charging load into the power distribution network become key.

At present, most documents develop electric vehicle charging load modeling research based on two dimensions of time and space. Some electric vehicles adopt a probability distribution function to obtain the charging starting time and daily driving mileage of the electric vehicle, and adopt a Monte Carlo simulation method to obtain the charging load of the electric vehicle according to the charging power and the charging time in the charging process; some methods adopt a queuing theory to model the flow of the charging station on the highway, so as to obtain the charging load in the charging station. The two methods are researched from the time dimension, but as a movable load, the charging load characteristic of the electric automobile is closely related to factors such as user travel behavior, battery endurance, traffic network topology and flow, and the coupling influence of a traffic network and a power distribution network needs to be comprehensively considered. Some methods adopt Origin-Destination (OD) analysis methods, obtain an OD matrix for simulating a vehicle driving path through back-stepping on related software through existing traffic data, and further determine the driving behavior of the electric vehicle based on the shortest path principle, but ignore the influence of road network flow on driving time, possibly resulting in very high time cost of the obtained travel path, and a simple energy consumption model of the electric vehicle. In order to simulate the driving behavior of a user in a road network more truly, the influence of factors such as a traffic road model, an electric vehicle model and user travel on the charging load distribution is comprehensively considered, and the time-space distribution characteristics of the charging load of the charging station at different spatial positions are accurately described.

On the other hand, in order to reduce the impact of the charging load of the electric vehicle on the power distribution network, many scholars absorb the charging load of the electric vehicle by introducing a distributed power generation technology into the power distribution network. Some charging stations introduce photovoltaic and energy storage systems, and establish an optimized scheduling model for reducing multiple targets such as electricity purchase cost and storage battery pack circulating electric quantity, but research objects only aim at the charging stations containing photovoltaic and energy storage, and influence of scheduling of the charging stations containing light storage on a power distribution network is not considered. Some models are established by a multi-target optimization scheduling model with minimum load fluctuation, maximum utilization rate of open renewable energy and maximum vehicle owner income under the constraint condition of considering access to a power grid, and an improved variable threshold optimization algorithm is used for coordinating energy exchange between the electric vehicle and the micro-grid, but the problem that the randomness of charging load fluctuation of the electric vehicle can affect the power distribution grid is not considered. Some electric vehicle users plan travel paths by using a road network 'balance model', a power distribution network robust scheduling model of a coupled traffic system is built, and the influence of charging load time sequence change on power distribution network scheduling is not considered.

Disclosure of Invention

The invention provides a power distribution network hybrid optimization scheduling method considering a light storage and fast charging integrated station, which aims to solve the technical problems in the prior art.

The technical scheme adopted by the invention for solving the technical problems in the prior art is as follows: a power distribution network hybrid optimization scheduling method considering a light storage and fast charging integrated station is characterized by establishing a road traffic network model, a driving speed and path model of an electric automobile and an electric automobile travel time model, establishing a charging load model of the fast charging integrated station based on the three models, and processing uncertainty of fast charging load by adopting a random optimization method; on the basis of a charging load model of the quick charging integrated station, simultaneously, an uncertain set is adopted to represent photovoltaic output and the range of original loads in a power distribution network, the minimum operation cost of the whole power distribution network to which the quick charging integrated station belongs is taken as a target function, and constraint conditions are set to establish a robust optimization scheduling model of the power distribution network of the quick charging integrated station; decomposing a power distribution network robust optimization scheduling model of the fast charging integration station into a main problem model and a sub problem model; the worst scene of the power distribution network tide operation is obtained by optimizing the photovoltaic output and the total time interval of the boundary value of the original load selection interval in the scheduling period; and solving the optimal solution of the main problem model and the sub problem model under the condition of meeting the convergence condition by adopting a column constraint generation algorithm to obtain a group of optimized scheduling strategy parameters consisting of main problem parameters and sub problem parameters.

Further, approximately simulating the corresponding initial travel time distribution and return travel time distribution of the electric vehicle by adopting a lognormal distribution and normal distribution probability model, and establishing an electric vehicle travel time model as shown in the following formula:

in the formula, μ and σ represent the mean and standard deviation of the travel time, respectively, and δ represents the travel time.

Further, the road traffic network model is established based on a classical graph theory method, an adjacency matrix D representing a road weight is adopted to describe the relation between traffic nodes and road sections, and the road traffic network model is established as shown in the following formula:

in the formula (d)_ijThe element of the ith row and the jth column in the matrix D represents the length of the corresponding road section; l_ijRepresents the length of the link (i, j); n is a radical of_TRepresenting the set of all traffic nodes in the road network.

Further, a running speed and path model of the electric automobile is established based on the shortest time in the whole course.

Further, the method for establishing the driving speed and path model of the electric automobile is as follows:

calculating the average passing time of each road section by the average speed of the electric vehicles on the passing road section as the basis of path optimization, wherein the relationship between the average speed of the electric vehicles in the road section b and the average passing time is shown as the following formula:

in the formula, v_b(t) is the average speed of all electric vehicles in the section b at the moment t; v. of_i(t) is the speed of the ith vehicle at time t; n is a radical of_bRepresenting the number of electric vehicles on the road section b at the moment t; s_b(t) is the average transit time of the section b at the moment t; l is_bRepresents the length of the section b;

when a user of the electric vehicle plans a travel path, the battery SOC and the energy consumption in the driving process are in the following linear relationship:

in the formula, SOC_dRepresenting the state of charge of the electric vehicle when the destination is predicted to be reached; SOC_oThe charge state of the electric automobile as a starting point; s_dRepresents a distance traveled; c_EVRepresenting the battery capacity of the electric automobile; e represents the predicted value of the energy consumption of the electric automobile in unit kilometer;

searching a path with shortest time in the whole course from the selectable paths for the electric vehicle user to travel by adopting an extent search traversal algorithm;

if the remaining capacity is not lower than the charging threshold value, the driving speed and the route model of the electric vehicle are as follows:

in the formula, o is a starting node; d is a destination node in the traffic road network; route_iRepresents the ith path from o to d; n is a radical of_totalRepresents Route_iA set of all road segments in;

if the remaining power is lower than the charging threshold value on the way, a proper charging station needs to be planned for charging, and the queuing time and the charging time of the user in the charging station need to be considered, so that the driving speed and the path model of the electric vehicle are as follows:

in the formula, PC_jThe j number public charging station in the area is represented;

representing the driving time corresponding to the ith route from the departure place o to the jth public charging station;

indicating the queuing time at the j-th public charging station;

represents the charging time at the j-th public charging station;

and represents the travel time corresponding to the ith route from the j-th public charging station to the destination d.

Further, the average speed of the electric automobile passing through the road section is obtained by adopting a cellular automata model, and the specific method comprises the following steps:

assuming that an electric vehicle is a cell with the length of n, wherein n is the length of the electric vehicle, the acceleration or braking time s of a driver is taken as the simulated time step length, and the maximum running speed of the vehicle is set as v_max(ii) a Let d_nRepresents the distance between the nth vehicle and the preceding vehicle, and is set as x_nThe position coordinate of each vehicle on the road section is shown, and v is set_nRepresenting the speed of each vehicle; setting the unit speed as a driving distance per second as a cell;

in the process from the t-th time step to the t + 1-th time step, the speed and the position of any vehicle on the road are updated in parallel according to the following rules:

1) slow start procedure, i.e. in a stationary state, if the front of the vehicle is a vacant cell, the vehicle has a certain probability p_staCarrying out accelerated starting; i.e. at v_nWhen (t) is 0, there is 1-p_staThe probability of (a), i.e. the speed during the slow start of the vehicle, is:

v_n(t+1)＝v_n(t)；

2) accelerate the process if v_n(t)<v_maxAnd increasing the running speed of the vehicle by one unit, namely the speed of the vehicle in the acceleration process is as follows:

v_n(t+1)＝min{v_n(t)+1,v_max}；

3) deceleration process, if d_n<v_n(t), the speed of the vehicle is reduced to d_nNamely, the speed of the vehicle in the deceleration process is as follows:

v_n(t+1)＝min{v_n(t+1),d_n}；

4) random slowing process, the speed of the vehicle being at a certain probability p_decDecreasing by one unit, i.e. the speed of the vehicle during stochastic slowing down, is:

v_n(t+1)＝max{v_n(t)-1,0}；

5) the position of the vehicle is updated as follows:

x_n(t+1)＝x_n(t)+v_n(t)。

further, an objective function of a robust optimization scheduling model of the power distribution network of the fast charging integration station is as follows:

in the formula:

C_gridrepresenting the power purchase cost of the power distribution network;

C_lossrepresenting line loss cost;

C_storagerepresenting the cost of energy storage charging and discharging;

Ω_ta set of scheduling moments for one operating cycle of a typical scene;

K_storagethe converted unit charge-discharge cost of the energy storage system is obtained;

Ω_essis an energy storage system set;

representing the discharge power of the mth energy storage system at the time t;

representing the charging power of the mth energy storage system at the moment t;

eta is the charge-discharge efficiency of the energy storage converter;

Ω_spresentation instrumentA set of scenes;

ω_sis the probability of occurrence of scene s;

the time-of-use electricity price at the time t is obtained;

representing the active power flowing into the power distribution network at the moment t of the scene s;

K_lossis the unit loss cost;

Ω_gridthe method comprises the steps of collecting all nodes of a power distribution network;

I_ij,s,trepresenting the current flowing from the node i to the node j in the power grid at the moment t of the scene s;

R_ijrepresenting the resistance value of branch ij.

Further, the set constraints include: the method comprises the following steps of power distribution network power flow operation restraint, energy storage device restraint and static reactive power compensation device restraint.

Further, the power distribution network flow operation constraint comprises:

u_min≤u_i,s,t≤u_max (s-7)；

0≤i_ij,s,t≤i_max (s-8)；

in the above formula:

i → j indicates that node i is the upstream node of node j;

j → k indicates that node j is an upstream node of node k;

i_ij,s,ta square value representing the magnitude of the current flowing from node i to node j at time t of scene s;

i_maxa square value representing an upper limit of the branch current;

u_i,s,tthe square value of the voltage amplitude of the node i at the moment t of the scene s is shown;

u_j,s,tcorresponding to the square value of the voltage amplitude of the node j at the moment t of the scene s;

u_minis the square value of the lower limit of the node voltage;

u_maxis the square of the upper limit of the node voltage;

P_ij,s,tthe active power flowing out from the node i to the node j at the moment t of the scene s;

Q_ij,s,tthe reactive power of the scene s flowing from the node i to the node j at the time t;

P_j,s,tthe net load active power of a node j at the moment t of a scene s;

Q_j,s,tthe net load reactive power of a node j at the moment t of a scene s;

P_jk,s,tthe active power flowing out from the node j to the node k at the moment t of the scene s;

Q_jk,s,tthe reactive power of the scene s flowing out from the node j to the node k at the moment t;

R_ijthe resistance value of the line between the node i and the node j;

X_ijthe reactance value of the line between the node i and the node j is obtained;

the active power of the original load of the node j at the moment t;

the reactive power of the original load of the node j at the moment t;

fast charging station charging load at a node j at the moment t of a scene s;

the reactive power is the reactive power sent by the static reactive power compensation device of the node j at the time t;

charging power of an energy storage system for a node j at the moment t;

the discharge power of the energy storage system is the node j at the moment t;

the active power of the photovoltaic output of the node j at the moment t.

Further, the energy storage device restraint comprises:

in the above formula:

charging power of an energy storage system for a node j at the moment t;

the discharge power of the energy storage system is the node j at the moment t;

the maximum charge-discharge power allowed by the node j energy storage converter is represented;

λ_j,tthe charging and discharging state of the energy storage battery at the node j at the time t is represented, discharging is represented when 1 is taken, and charging is represented when 0 is taken;

t represents the total number of scheduling periods;

storing the residual capacity of the node j at the initial scheduling moment;

represents the energy storage capacity of node j;

allowable SOC upper limit during operation for energy storage；

The lower limit of the allowable SOC during the operation process of the stored energy;

eta is the charge-discharge efficiency of the energy storage converter;

Δ t represents the simulation step size.

Further, a Monte Carlo simulation method is adopted to generate a charging load scene of a fast charging station in the power distribution network, a scene clustering method is adopted to obtain a typical scene, and the worst scene of tidal current operation of the power distribution network is further obtained.

Further, the following uncertain sets are used to characterize the range of the photovoltaic output and the original load in the distribution network:

in the formula (I), the compound is shown in the specification,

the active power of the original load of the node j at the time t,

the active power of photovoltaic output of a node j at the time t;

and

corresponding to the predicted values of the photovoltaic output and the original load power;

and

correspondingly representing the fluctuation deviation of the photovoltaic output and the original load power; gamma-shaped^PVTaking values of uncertain regulation parameters introduced by photovoltaic outputThe integral number in the range of 0-T represents the total time interval of the minimum value or the maximum value of the photovoltaic output fluctuation interval in the scheduling period; gamma-shaped^LThe method comprises the steps that uncertain adjusting parameters are introduced aiming at original load power, the value range is an integer from 0 to T, and the total number of time periods when the original load power takes the minimum value or the maximum value of a fluctuation interval in a scheduling period is represented; t represents the total number of scheduling periods; gamma-shaped^PVAnd Γ^LThe conservative property for adjusting the optimal solution is that the scheme obtained by the method is more conservative when the value is larger, and conversely, the scheme is more risky.

Further, the method for solving the optimal solution of the main problem model and the sub problem model under the condition of meeting the convergence condition by adopting the column constraint generation algorithm comprises the following steps:

step I: setting x as a parameter to be optimized of the main problem model; setting f (x) as an optimal value expression of the sub-problem model, and setting y and v as parameters to be optimized of the sub-problem model, wherein v represents the worst scene; setting the lower bound of the overall operation cost of a power distribution network to which the quick charging integrated station belongs as LB ═ infinity, and setting the upper bound of the overall operation cost as UB ═ infinity; setting the iteration times as l, and setting the initial value of l as 1; given an initial worst scenario value v₁；

Step II: setting the value of the worst scene as v in the first iteration_l: according to the worst scene v_lSolving the main problem model to obtain the optimal solution of x, and setting

The optimized value of the first iteration of the x parameter is obtained;

the optimization result of the first iteration of the main problem model is obtained; order to

Step III: will be provided with

Substituting the sub-problem model to obtain the optimal value of the sub-problem model

And worst case scenario

Order to

Step IV: judging whether a convergence condition is met, if UB-LB is less than epsilon, wherein epsilon is a small enough convergence threshold value which is set in advance, indicating that an optimal solution is obtained, and turning to the step VI; otherwise, turning to the step V;

step V: adjusting the sub-problem constraints and sub-problem parameter values, let l be l +1, let

Returning to the step II;

step VI: and finishing the iteration.

The invention has the advantages and positive effects that: on the basis of modeling of the charging load, the uncertainty of the fast charging load, the photovoltaic output and the original load of the power distribution network is considered, and a random optimization and robust optimization method is combined to provide a power distribution network random/robust hybrid optimization scheduling method considering the light storage and fast charging integrated station, so that the negative influence of the fast charging load on the power distribution network is reduced. The decision of the energy storage system of the quick charging station under the uncertain environment can be flexibly realized, so that the reliable and economic operation of the power distribution network is guaranteed.

Drawings

Fig. 1 is a schematic diagram of a road network-grid coupling system.

Fig. 2 is a schematic time-of-use electricity price diagram of a power distribution network transaction.

Fig. 3 is a diagram of the distribution of the starting travel time of the electric automobile.

FIG. 4 is a diagram of a return time distribution of an electric vehicle such as an electric vehicle.

Fig. 5 is a diagram of distribution of starting travel time of taxi electric vehicles.

Fig. 6 is a schematic diagram of the charging load condition of the first energy storage quick charging station.

Fig. 7 is a schematic diagram of the charging load condition of the second energy storage quick charging station.

Fig. 8 is a schematic diagram of the charging load condition of the third energy storage quick charging station.

Fig. 9 is a schematic diagram of the charging load condition of the fourth energy storage quick charging station.

Fig. 10 is a schematic diagram of photovoltaic output of a distribution network.

Fig. 11 is a schematic diagram of the original load condition of the distribution network.

Fig. 12 is a schematic diagram of the load and energy storage charging and discharging power conditions of the power distribution network in the scene of the maximum uncertain budget value.

Detailed Description

For further understanding of the contents, features and effects of the present invention, the following embodiments are enumerated in conjunction with the accompanying drawings, and the following detailed description is given:

in the present invention, the English abbreviation is explained as follows:

SOC: state of charge of the electric vehicle.

Veh: provided is an electric automobile.

Referring to fig. 1 to 12, a hybrid optimization scheduling method for a power distribution network in a light storage and rapid charging integrated station is provided, which includes establishing a road traffic network model, a driving speed and path model of an electric vehicle, and an electric vehicle travel time model, establishing a charging load model of the rapid charging integrated station based on the three models, and processing uncertainty of rapid charging load by using a random optimization method; on the basis of a charging load model of the quick charging integrated station, simultaneously adopting an uncertain set to represent photovoltaic output and the range of original load in the power distribution network, setting a constraint condition to establish a robust optimization scheduling model of the power distribution network of the quick charging integrated station by taking the minimum overall operation cost of the power distribution network to which the quick charging integrated station belongs as a target function; decomposing a power distribution network robust optimization scheduling model of the fast charging integration station into a main problem model and a sub problem model; the worst scene of the power distribution network tide operation is obtained by optimizing the photovoltaic output and the total time interval of the boundary value of the original load selection interval in the scheduling period; and solving the optimal solution of the main problem model and the sub problem model under the condition of meeting the convergence condition by adopting a column constraint generation algorithm to obtain a group of optimal scheduling strategy parameters consisting of main problem parameters and sub problem parameters.

Preferably, the log-normal distribution and normal distribution probability model can be used to approximately simulate the corresponding initial trip time and return trip time distribution of the electric vehicle, and the trip time model of the electric vehicle is established as shown in the following formula:

Preferably, the road traffic network model may be established based on a classical graph theory method, which may describe a relationship between traffic nodes and road segments by using an adjacent matrix D representing a road weight, and the established road traffic network model may be represented as follows:

Preferably, the running speed and path model of the electric automobile can be established based on the shortest time in the whole course.

Preferably, the method for establishing the driving speed and path model of the electric vehicle may be as follows:

the average passing time of each road section can be calculated according to the average speed of the electric automobile on the passing road section, and the average speed and the average passing time of the electric automobile in the road section b are taken as the basis of path optimization, and the relationship between the average speed and the average passing time of the electric automobile in the road section b is shown as the following formula:

when a user of the electric vehicle plans a travel path, the battery SOC and the energy consumption in the driving process are in a linear relationship as follows:

an extent search traversal algorithm can be adopted to search a path with the shortest time in the whole process from the selectable paths for the electric vehicle user to go out;

if the remaining capacity is not lower than the charging threshold, the driving speed and the route model of the electric vehicle may be as follows:

if the remaining power is lower than the charging threshold value on the way, a proper charging station needs to be planned for charging, and the queuing time and the charging time of the user in the charging station need to be considered, so that the driving speed and the route model of the electric vehicle can be as follows:

indicating the queuing time at the j-th public charging station;

represents the charging time at the j-th public charging station;

Preferably, the cellular automaton model can be used to obtain the average speed of the electric vehicle passing through the road section, and the specific method is as follows:

an electric automobile is assumed to be a cellular with the length of n, the length of n is the length of the electric automobile, the acceleration or braking time s of a driver is taken as the simulated time step length, and the maximum speed of the vehicle is set as v_max(ii) a Let d_nRepresents the distance between the nth vehicle and the preceding vehicle, and is set as x_nThe position coordinate of each vehicle on the road section is shown, and v is set_nRepresenting the speed of each vehicle; setting the unit speed as a driving distance per second as a cell;

in the process from the t-th time step to the t + 1-th time step, the speed and the position of any vehicle on the road can be updated in parallel according to the following rules:

v_n(t+1)＝v_n(t)；

v_n(t+1)＝min{v_n(t)+1,v_max}；

v_n(t+1)＝min{v_n(t+1),d_n}；

v_n(t+1)＝max{v_n(t)-1,0}；

5) the position of the vehicle is updated as follows:

x_n(t+1)＝x_n(t)+v_n(t)。

according to the road traffic network model, the driving speed and path model of the electric automobile and the traveling time model of the electric automobile introduced above, the traveling situation and the charging situation of the electric automobile in a certain area can be simulated, and the specific simulation flow is as follows:

step a 1: according to an electric vehicle travel time model reflecting the distribution of the electric vehicle starting travel time and the return travel time, starting travel time t0 and return travel time tl of all electric vehicles and taxis in the area are generated, a starting departure place o and a corresponding destination d are randomly generated, and the starting time t is 0.

Step a 2: the current time t is t + 1. According to the relation between the average speed of the electric automobile in the road section and the average passing time, updating the average passing speed of all the road sections in the area in real time, calculating the passing time of the road section according to the length of the road section, and updating the queuing time of each charging station.

Step a 3: counting and compiling an electric vehicle set going out at the current moment, setting the electric vehicle set going out at the current moment as Veh1, and setting a vehicle number j to be 0; counting and compiling an electric vehicle set running at the current moment; let the set of electric vehicles running at the present time be Veh2, and the vehicle number i be 0.

Step a 4: the current vehicle number j ═ j + 1. According to the trip demand of the vehicle at the current moment, an extent search traversal algorithm is adopted to find a path with the shortest time in the whole process, plan a trip path and estimate the energy consumption condition of the trip. If the battery residual capacity is lower than the SOC charging threshold value of the user in the driving process, selecting a quick charging station according to the shortest time path by integrating the driving time of all driving paths, the waiting time and the charging time of charging stations, searching and planning the optimal charging path again, and adding the vehicle j into the vehicle set Veh2 driven at the current time in the next step; if it is predicted that the remaining battery capacity will not be lower than the user SOC charging threshold while driving, the vehicle j is directly added to the driving vehicle set Veh 2.

Step a 5: a determination is made as to whether j is greater than Veh 1. If yes, generating a vehicle set Veh2 driven at the current time, wherein the vehicle number i is 0; otherwise go to step a4 to continue execution.

Step a 6: the current vehicle number i ═ i + 1. Judging whether the vehicle reaches the destination at the current moment, if so, generating a next trip demand, and executing the step again; and if the destination is not reached at the current moment, determining the state of the electric vehicle at the current moment and updating the running information of the vehicle.

Step a 7: judging whether i is larger than Veh2, if so, continuing to execute the step a 8; otherwise, the operation is continued by switching to the step a 6.

Step a 8: judging whether the current time t is greater than a simulation period, wherein the simulation period is the total time step, the simulation step is 1 second, the simulation period is one day, the total time step is 86400 seconds, and if yes, continuing to execute the step a 9; otherwise go to step a 2.

Step a 9: and (5) after the simulation in one day is finished, counting the charging load condition of each quick charging station in the area in one day.

Preferably, an objective function of the robust optimization scheduling model of the power distribution network of the fast charging integration station can be as follows:

in the formula:

C_gridrepresenting the power purchase cost of the power distribution network;

C_lossrepresenting line loss cost;

C_storagerepresenting the cost of energy storage charging and discharging;

Ω_ta set of scheduling moments for one operating cycle of a typical scene;

Ω_essis an energy storage system set;

eta is the charge-discharge efficiency of the energy storage converter;

Ω_srepresents a collection of all scenes;

ω_sis the probability of occurrence of scene s;

the time-of-use electricity price at the time t is obtained;

K_lossis the unit loss cost;

R_ijrepresenting the resistance value of branch ij.

Preferably, the set constraints may include: the method comprises the following steps of power distribution network power flow operation restraint, energy storage device restraint and static reactive power compensation device restraint.

Preferably, the power distribution network flow operation constraints may include:

u_min≤u_i,s,t≤u_max (s-7)；

0≤i_ij,s,t≤i_max (s-8)；

in the above formula:

i → j indicates that node i is the upstream node of node j;

j → k indicates that node j is an upstream node of node k;

i_maxa square value representing an upper limit of the branch current;

u_minis the square value of the lower limit of the node voltage;

u_maxis the square of the upper limit of the node voltage;

P_j,s,tthe net load active power of a node j at the moment t of a scene s;

Q_j,s,tthe net load reactive power of a node j at the moment t of a scene s;

R_ijthe resistance value of the line between the node i and the node j;

the active power of the original load of the node j at the moment t;

the reactive power of the original load of the node j at the moment t;

fast charging station charging load at a node j at the moment t of a scene s;

charging power of an energy storage system for a node j at the moment t;

the discharge power of the energy storage system is the node j at the moment t;

the active power of the photovoltaic output of the node j at the moment t.

Preferably, the energy storage device restraint may comprise:

in the above formula:

charging power of an energy storage system for a node j at the moment t;

the discharge power of the energy storage system is the node j at the moment t;

t represents the total number of scheduling periods;

storing the residual capacity of the node j at the initial scheduling moment;

represents the energy storage capacity of node j;

the allowable SOC upper limit in the operation process of the stored energy is defined;

eta is the charge-discharge efficiency of the energy storage converter;

Δ t represents the simulation step size.

Preferably, a Monte Carlo simulation method can be adopted to generate a charging load scene of a fast charging station in the power distribution network, a scene clustering method can be adopted to obtain a typical scene, and the worst scene of the tidal current operation of the power distribution network can be further obtained.

Preferably, the range of photovoltaic output and the original load within the distribution network can be characterized using the following indeterminate set:

in the formula (I), the compound is shown in the specification,

the active power of the original load of the node j at the time t,

the active power of photovoltaic output of a node j at the time t;

and

and

correspondingly representing the fluctuation deviation of the photovoltaic output and the original load power; gamma-shaped^PVAiming at uncertain adjusting parameters introduced by photovoltaic output, the value range is an integer between 0 and TThe total number of periods of the minimum value or the maximum value of the photovoltaic output fluctuation interval in the scheduling period is represented; gamma-shaped^LThe method comprises the steps that uncertain adjusting parameters are introduced aiming at original load power, the value range is an integer from 0 to T, and the total number of time periods when the original load power takes the minimum value or the maximum value of a fluctuation interval in a scheduling period is represented; t represents the total number of scheduling periods; gamma-shaped^PVAnd Γ^LThe conservative property for adjusting the optimal solution is that the scheme obtained by the method is more conservative when the value is larger, and conversely, the scheme is more risky.

Preferably, the method for solving the optimal solution of the main problem model and the sub problem model under the condition of meeting the convergence condition by adopting the column constraint generation algorithm may comprise the following steps:

The optimized value of the first iteration of the x parameter is obtained;

Step III: will be provided with

And worst case scenario

Order to

Returning to the step II;

step VI: and finishing the iteration.

The working process and working principle of the present invention are further explained by a preferred embodiment of the present invention as follows:

the invention relates to a power distribution network hybrid optimization scheduling method considering a light storage and fast charging integrated station, which comprises the steps of establishing a road traffic network model, a driving speed and path model of an electric automobile and an electric automobile travel time model, establishing a charging load model of the fast charging integrated station based on the three models, and processing uncertainty of fast charging load by adopting a random optimization method; on the basis of a charging load model of the quick charging integrated station, simultaneously adopting an uncertain set to represent photovoltaic output and the range of original load in the power distribution network, setting a constraint condition to establish a robust optimization scheduling model of the power distribution network of the quick charging integrated station by taking the minimum overall operation cost of the power distribution network to which the quick charging integrated station belongs as a target function; decomposing a power distribution network robust optimization scheduling model of the fast charging integration station into a main problem model and a sub problem model; the worst scene of the power distribution network tide operation is obtained by optimizing the photovoltaic output and the total time interval of the boundary value of the original load selection interval of the power distribution network in the scheduling period; and solving the optimal solution of the main problem model and the sub problem model under the condition of meeting the convergence condition by adopting a column constraint generation algorithm to obtain a group of optimized scheduling strategy parameters consisting of main problem parameters and sub problem parameters.

The invention relates to a power distribution network hybrid optimization scheduling method considering a light storage and fast charging integrated station, which specifically comprises the following steps:

1. and establishing a charging load model of the quick charging station.

a) And establishing an electric automobile travel time model.

Approximately simulating the corresponding initial travel time and return travel time distribution of the electric automobile by adopting a lognormal distribution and normal distribution probability model:

in the formula, μ and σ represent the mean and standard deviation of the travel time, respectively.

Considering the energy consumption modeling of the electric automobile, the vehicle-mounted battery of the electric automobile discharges through the direct current side, is converted into alternating current through the vehicle-mounted inverter, and is supplied to the permanent magnet alternating current motor, so as to drive a mechanical device (a clockwork spring, a tire and the like) to move. The loss of the electric automobile running on the road mainly comprises four parts: tire losses, air drag losses, active system losses and auxiliary losses. The loss is calculated as follows:

F_fric+F_aero＝A₁+A₂v+A₃v² (3)。

F_acce＝ma (4)。

in the formula, F_fric、F_aeroAnd F_acceRespectively representing the tire resistance, the air resistance and the force required by the active system in an acceleration state; v is the velocity; a. the₁、A₂、A₃Respectively resistance parameters; m is the weight of the electric automobile; a represents acceleration; m is the torque required at the motor; m_CFIs the no-load moment after conversion; r is the tire radius; n is_gAnd η_gRespectively representing the transmission coefficient ratio and the transmission efficiency; p_dcRepresenting the active power of the direct current side of the battery; eta_mIs the efficiency of the motor; PF (particle Filter)_mIs the power factor of the motor; eta_invThe inverter transmission efficiency; p_auxIndicating the auxiliary losses.

b) And establishing a traffic network model.

Referring to a classical graph theory method, a tie matrix D representing road weight is adopted to describe the relation between traffic nodes and road sections. Element d in the matrix_ijLength of the corresponding link is represented:

in the formula I_ijRepresents the length of the link (i, j); n is a radical of_TRepresenting the set of all traffic nodes in the road network.

c) And establishing an electric automobile traveling process model.

c-1) establishing an electric automobile travel path model.

Calculating the average passing time of each road section according to the average speed of the electric vehicles on the road section, wherein the average passing time is used as the basis of path optimization, and the average passing speed and the passing time of the road section b are obtained according to the following formula:

in the formula, v_b(t) is the average speed of all electric vehicles in the section b at the moment t; v. of_i(t) is the speed of the ith vehicle at time t; n is a radical of_bRepresenting the number of electric vehicles on the road section b at the moment t; s_b(t) is the average transit time of the section b at the time t; l is_bIndicating the length of the section b.

Meanwhile, when a user of the electric vehicle plans a travel path, the battery SOC (State of charge) and the energy consumption in the driving process are in the following linear relationship:

in the formula, SOC_dRepresenting the state of charge of the electric vehicle when the destination is predicted to be reached; SOC_oThe state of charge of the electric vehicle as a starting point; s_dRepresents a distance traveled; c_EVRepresents the battery capacity of the electric vehicle. And e represents a predicted value of the energy consumption of the electric automobile in a unit of public.

When the travel path of the electric vehicle user is simulated, the path with the shortest time in the whole range is searched by adopting an extent search traversal algorithm. If the situation that the residual capacity is lower than the charging threshold value does not occur on the way, solving the shortest time-using path as follows:

in the formula, o is a starting node; d is a destination node in the traffic road network; route_iRepresents the ith path from o to d; n is a radical of_totalRepresents Route_iZhongshiThere is a collection of road segments.

If the remaining capacity is lower than the charging threshold value, a proper charging station needs to be planned for charging, and the queuing time and the charging time of the user in the charging station need to be considered:

indicating the queuing time at the j-th public charging station;

represents the charging time at the j-th public charging station;

And c-2) modeling the travel process of the electric automobile.

And introducing a cellular automaton model to simulate the traveling process of the electric automobile on the road in detail. Assuming that a car is 5 meters in length, which is the length occupied by a car equivalent when a road is heavily congested, every 5 meters on a road section is taken as a cell. Assuming that the reaction (acceleration or braking) time of the driver is 1s, as the simulation time step, considering that the maximum speed of the automobile running on the urban road generally does not exceed 60km/h, the automobile can not move more than 3 cells within one simulation step, namely the maximum passing speed v of the automobile running_maxIs 3 cells/s. d_nRepresenting the distance between the nth vehicle and the preceding vehicle, and the motion state of each vehicle is changed by twoLocation x of measurement_nAnd velocity v_nDescribing, in the process of time step t → t +1, the speed and position of all vehicles on the road are updated in parallel according to the following rules:

and (1) slowly starting the process. I.e. in a stationary state, if the front is a vacant cell, the vehicle has a certain probability p_staAn accelerated start is performed. I.e. at v_nWhen (t) is 0, there is 1-p_staProbability of (c):

v_n(t+1)＝v_n(t) (15)。

② the process is accelerated. If v is_n(t)<v_maxThen the travel speed of the vehicle is increased by one unit, namely:

v_n(t+1)＝min{v_n(t)+1,v_max} (16)。

and thirdly, a deceleration process. If d is_n<v_n(t), the speed of the vehicle is reduced to d_nNamely:

v_n(t+1)＝min{v_n(t+1),d_n} (17)。

and fourthly, randomly slowing down. As long as the vehicle is travelling on the road, its speed will be at a certain probability p_decDecreasing by one unit. This takes into account the effect that other uncertainties may have on the reduction in vehicle speed, namely:

v_n(t+1)＝max{v_n(t)-1,0} (18)。

position updating. The electric automobile updates the position according to the speeds obtained by the equations (12) to (14), namely:

x_n(t+1)＝x_n(t)+v_n(t) (19)。

according to the road traffic model, the electric automobile model and the user travel model introduced above, the travel situation and the charging situation of the electric automobile in a certain area can be simulated, and the specific simulation flow is as follows:

step 1: generating starting travel time t0 and return time tl of all electric vehicles and taxis in the area according to the formulas (1) and (2), and randomly generating a starting place o and a corresponding destination d, wherein the starting time t is 0;

step 2: the current time t is t + 1. Updating the average passing speed of all road sections in the area in real time through the formula (10) and the formula (11) according to the road condition at the current moment, calculating the passing time of the road sections according to the length of the road sections, and updating the queuing time of each charging station;

and step 3: counting the electric automobile Veh1 going out at the current moment, wherein the automobile number j is 0;

and 4, step 4: the current vehicle number j ═ j + 1. According to the trip demand of the vehicle at the current moment, a trip path is planned through a formula (13) and the energy consumption condition of the trip is estimated through a formula (12). If the estimated battery residual capacity is lower than the user SOC charging threshold value in the driving process, selecting a quick charging station according to the formula (14), searching and planning an optimal charging path again, and adding the vehicle j into a vehicle set Veh2 driven at the current moment in the next step; if the fact that the remaining battery capacity is lower than the SOC charging threshold of the user does not occur in the process of driving, directly adding the vehicle j into the driving vehicle set Veh 2;

and 5: a determination is made as to whether j is greater than Veh 1. If yes, generating a vehicle set Veh2 driven at the current time, wherein the vehicle number i is 0; otherwise, go to step 4 to continue execution;

step 6: the current vehicle number i ═ i + 1. Judging whether the vehicle reaches the destination at the current moment, if so, generating a next trip demand, and executing the step again; if the destination is not reached at the current moment, determining the state of the electric vehicle at the current moment according to the formula (15) to the formula (19), and updating the running information of the vehicle;

and 7: judging whether i is larger than Veh2, if so, executing step 8; otherwise, the operation is continued in the step 6.

And 8: judging whether the current time t is greater than a simulation period, wherein the simulation period is a total time step, the simulation step is 1 second, the simulation period is one day, and therefore the total time step is 86400 seconds, if yes, turning to the step 9; otherwise, go to step 2.

And step 9: and (5) after the simulation in one day is finished, counting the charging load condition of each quick charging station in the area in one day.

2. And establishing a power distribution network random optimization scheduling model considering the light storage and fast charging integrated station.

a) And processing the uncertainty of the fast charging load by adopting a random optimization method. According to the rapid charging modeling process, a large number of rapid charging load scenes in the region can be generated by adopting a Monte Carlo simulation method, and then a typical scene with concentrated original scenes is obtained by adopting a scene clustering method.

b) Assuming that a fast charging station, a photovoltaic system, an energy storage system and the like in the power distribution network are invested by a power distribution network operator, the objective function of optimizing the scheduling is to minimize the day-ahead operation cost of the power distribution network operator:

in the formula, C in the first row_gridIndicating the electricity purchase cost of the distribution network, C_lossRepresents the line loss cost, C_storageRepresenting the energy storage charge-discharge cost. Second row middle omega_tSet of scheduling instants for a typical scenario-one operating cycle, K_storageThe unit charge-discharge cost, omega, of the energy storage system after conversion_essIs a set of energy storage systems, and is,

representing the discharge power of the mth energy storage system at time t,

and (3) representing the charging power of the mth energy storage system at the time t, wherein eta is the charging and discharging efficiency of the energy storage converter. Third row middle omega_sSet representing all scenes, ω_sIs the probability that the scene s occurs,

the time-of-use electricity price at the time t is obtained;

representing the active power flowing into the distribution network at time t of scene s. In the fourth row K_lossIn terms of the cost per unit of loss of the network,Ω_gridfor all node sets of the distribution network, I_ij,s,tRepresenting the current of a node i to a node j in the grid at time t of a scene s, R_ijRepresenting the resistance value of branch ij.

c) Constraint conditions

u_min≤u_i,s,t≤u_max (27)。

0≤i_ij,s,t≤i_max (28)。

Equations (21) - (28) are power flow constraints; equations (29) - (33) are energy storage and SVC device constraints.

In the above formula:

i → j indicates that node i is the upstream node of node j;

j → k indicates that node j is an upstream node of node k;

i_maxa square value representing an upper limit of the branch current;

u_minis the square value of the lower limit of the node voltage;

u_maxis the square of the upper limit of the node voltage;

P_j,s,tthe net load active power of a node j at the moment t of a scene s;

Q_j,s,tthe net load reactive power of a node j at the moment t of a scene s;

Q_jk,s,tfor scene s at time tThe reactive power flowing out from the node j to the node k;

R_ijthe resistance value of the line between the node i and the node j;

the active power of the original load of the node j at the moment t;

the reactive power of the original load of the node j at the moment t;

fast charging station charging load at a node j at the moment t of a scene s;

charging power of an energy storage system for a node j at the moment t;

the discharge power of the energy storage system is the node j at the moment t;

the active power of the photovoltaic output of the node j at the moment t.

t represents the total number of scheduling periods;

storing the residual capacity of the node j at the initial scheduling moment;

represents the energy storage capacity of node j;

eta is the charge-discharge efficiency of the energy storage converter;

Δ t represents the simulation step size.

Representing the maximum compensation capacity of the static var compensation apparatus.

3. And establishing a power distribution network random/robust hybrid optimization scheduling model considering the light storage and fast charging integrated station.

In a power distribution network comprising a light storage and quick charge integrated station, the randomness of quick charge load is considered, and the original load of a photovoltaic power generation and power distribution network has obvious uncertainty in the daily operation process. The method comprises the following steps of (1) adopting a processing variable uncertainty box type uncertain set in robust optimization to characterize the range of photovoltaic output and original load in a power distribution network:

PV represents photovoltaic, and L represents the original load of the power distribution network; introduction of

And

correspondingly representing uncertain parameters of photovoltaic output and original load power of the power distribution network; wherein the content of the first and second substances,

the active power of the original load of the node j at the time t,

the active power of photovoltaic output of a node j at the time t;

and

and

correspondingly representing the fluctuation deviation of the photovoltaic output and the original load power; gamma-shaped^PVThe photovoltaic output power control method comprises the steps of (1) aiming at uncertain adjustment parameters introduced by photovoltaic output power, wherein the value range is an integer from 0 to T, and the total number of time periods of the minimum value or the maximum value of a photovoltaic output power fluctuation interval in a scheduling period is represented; gamma-shaped^LThe method comprises the steps that uncertain adjustment parameters are introduced aiming at original load power, the value range is an integer from 0 to T, and the total number of time segments of the original load power which takes the minimum value or the maximum value of a fluctuation interval in a scheduling period is represented; t represents the total number of scheduling periods; gamma-shaped^PVAnd Γ^LConservation for adjusting optimal solution, the larger the value isThe more conservative the resulting solution, and conversely, the more risky the solution.

On the basis of equation (34), equation (20) can be converted into a stochastic/robust hybrid optimization scheduling model as follows:

in the formula, omega_tSet of scheduling instants for a typical scenario-one operating cycle, K_storageThe unit charge-discharge cost of the energy storage system after conversion is omega_essIs a set of energy storage systems, and is,

representing the discharge power of the mth energy storage system at time t,

the charging power of the mth energy storage system at the moment t is represented, and eta is the charging and discharging efficiency of the energy storage converter; omega_sSet representing all scenes, ω_sIs the probability that the scene s occurs,

the time-of-use electricity price at the time t is obtained;

representing the active power of a scene s flowing into a power distribution network at the moment t; k_lossIs unit loss cost, omega_gridFor all node sets of the distribution network, I_ij,s,tRepresenting the current of a node i to a node j in the grid at time t of a scene s, R_ijRepresenting the resistance value of branch ij.

From the above equation, it can be seen that the model changes from single-phase to two-phase after considering the uncertain set of photovoltaic output and load power. The constraint conditions are formula (21) to formula (33). For convenience of explanation hereinafter, a compact form of the model is given as follows:

wherein the first stage optimization variable is set

Comprises an energy storage charging and discharging state column vector lambda and an energy storage charging and discharging power column vector

And

and the reactive power compensation device SVC outputs the reactive power column vector

The second stage optimization variable is y ═ P_ij,s,t，Q_ij,s,t，u_i,s,t，i_ij,s,t]And

the method is used for describing the power distribution network tide optimization solution and the uncertain variables of photovoltaic output and load in each scene. Wherein, λ is the energy storage charge-discharge state column vector;

the discharge power of the energy storage system is the node j at the moment t;

charging power of an energy storage system for a node j at the moment t;

the reactive power is the reactive power sent by the static reactive power compensation device of the node j at the time t; p_ij,s,tAnd Q_ij,s,tCorresponding to the active power and reactive power flowing from the node i to the node j at the moment t of the scene sPower; u. of_i,s,tThe square value of the voltage amplitude of the node i at the moment t of the scene s is shown; i.e. i_ij,s,tRepresenting the square of the magnitude of the current flowing from node i to node j at time t of scene s,

the active power of the original load of the node j at the time t,

the active power of photovoltaic output of a node j at the time t; c. h is a coefficient matrix corresponding to the objective function; A. c, D, E, G, R, g is a coefficient matrix corresponding to the variables under the constraint; a. r is a constant column vector. The first row of constraints in equation (37) corresponds to equations (29) to (33); Δ (x, v) represents the feasible domain of the second stage variable y given a set of x and v, where the first row constrains the corresponding equations (25) - (26), the second row constrains the corresponding equations (21) - (23) and (27) - (28), and the third row constrains the corresponding equation (24).

Solving the problem by adopting a column constraint generation algorithm, wherein the specific process is as follows:

a) the main problem model is in the following specific form:

in the formula, psi represents the operation cost of the power distribution network except the energy storage cost, and l is the current iteration number; x is the number of_lThe solution of the main problem model in the first iteration; y is_lThe solution of the subproblem model after the first iteration; v. of_lThe value of the uncertain variable v in the worst scene obtained after the first iteration is obtained; k is the number of iterations.

b) The sub-problem model is in the following specific form:

in the formula, x^*And carrying out solving on the sub-problem for the optimization result of the main problem in the first stage.

c) Dual form of subproblems:

can be set as gamma according to strong dual theory₁Corresponding to Ey ═ Dx^*+ v dual variable, which may be set to γ₂Corresponding to Ay ≧ Cx^*A dual variable of + a; can set omega_i、

Corresponding to less than or equal to g in the formula of Gy | | | |^TA dual variable of y; converting the sub-problem model into a dual form as follows;

when the maximum value is obtained by the equation (40), the value of the uncertain variable v should be the boundary of the fluctuation interval described by the equation (14). In addition, for the problem to be studied, when the fast charging load reaches the maximum value, the operation cost of the power distribution network is higher, and the operation cost better conforms to the definition of the worst scene, so the formula (11) can be rewritten into the following form:

And

the active power of the original load of the node j at the time t,

the active power of photovoltaic output of a node j at the time t;

and

corresponding to the preset values of photovoltaic output and original load power of the power distribution network;

and

correspondingly representing the fluctuation deviation of the photovoltaic output and the original load power of the power distribution network; gamma-shaped^PVThe numerical control method comprises the steps of (1) aiming at uncertain adjustment parameters introduced by photovoltaic output, wherein the numerical range is an integer from 0 to T, and the total number of time periods of the photovoltaic output taking the minimum value or the maximum value of a fluctuation interval in a scheduling period is represented; gamma-shaped^LThe method comprises the steps of introducing uncertain adjusting parameters aiming at the original load power of the power distribution network, wherein the value range of the uncertain adjusting parameters is an integer from 0 to T, and the total number of time periods when the original load power of the power distribution network takes the minimum value or the maximum value of a fluctuation interval in a scheduling period is represented; t represents the total number of scheduling periods;

and

is an introduced auxiliary variable.

Substituting the uncertain variable expression in equation (41) into equation (40), and introducing the auxiliary variable B' can transform the subproblem into the following mixed integer second order cone programming model:

in the formula (I), the compound is shown in the specification,

represents the even-to-even variable gamma₂The upper boundary of (2) can be calculated as a sufficiently large positive real number.

d) The column constraint generation algorithm solves the model. After the derivation and conversion, the two-stage robust model is finally decoupled into a main problem equation (38) and a sub problem equation (42), and then can be solved by using a column constraint generation algorithm, and the specific steps are as follows:

step 1: a group of v values is given as an initial worst scene, a lower limit LB ═ infinity, an upper limit UB ∞ infinity and an iteration number l ═ 1 of the operation cost are set;

step 2: according to the worst scene v_lSolving the main problem to obtain the optimal solution

And

obtained as a major problem

Value as a new lower bound

And step 3: will be provided with

And worst case scenario

Order to

And 4, step 4: if UB-LB<Epsilon, wherein epsilon is a small enough convergence threshold set in advance, which indicates that the best solution has been obtained, and the iteration is stopped; otherwise, adjusting the parametersAnd constraint, increasing variable y_l+1And the following constraints:

let l be l +1, return to step 2 until the algorithm converges.

The invention adopts a traffic network with 13 nodes and an adjusted IEEE33 node power distribution network coupling system for verification, wherein the area network of the power distribution network marked with the number 1 is a first area network, the area network of the power distribution network marked with the number 2 is a second area network, the area network of the power distribution network marked with the number 3 is a third area network, and a system topological diagram is shown in figure 1. The simulation duration is 1 hour, and the time-of-use electricity price of the power distribution network transaction is shown in fig. 2. The four fast charging stations are respectively built on four

traffic nodes

6, 7, 9 and 11 and are respectively a first energy storage fast charging station, a second energy storage fast charging station, a third energy storage fast charging station and a fourth energy storage fast charging station; the quantity of the fast charging piles of the first to fourth energy storage fast charging stations is 20, 10, 20 and 15 in sequence, the fast charging power of each fast charging pile is 45kW, and the fast charging efficiency is 95%. Travel parameters of various electric vehicles are shown in fig. 3 to 5: the electric automobile initial trip time follows logarithmic normal distribution, the mean value is 2.18, the standard deviation is 0.30, the return trip time follows standard normal distribution, the mean value is 16.79, and the standard deviation is 3.53; the initial trip time of the taxis is distributed according to the piecewise lognormal distribution, the taxis which are distributed according to 1 section are assumed to be public taxis (1 type), the mean value is-1.68, the standard deviation is 1.41, the return time is 21 to 23 points, the taxis which are distributed according to 2 sections are private taxis (2 types), the mean value is 1.90, the standard deviation is 0.36, and the return time is 19 to 23 points.

There are 2500 electric cars and 700 taxis (both electric cars, wherein the taxis include 150 class 1 taxis and 550 class 2 taxis) in the area, the vehicle type is the electric car of a Nissan Leaf, and when a path is planned for the electric car, e is 0.15 kWh/km. The battery capacity was 24kWh, and the coefficients of the energy consumption model are shown in table 1. Assuming that the SOC range of the vehicle in the initial trip is 0.8-0.9, the charging demand threshold is 0.2-0.3, and the upper limit of the charging SOC isIs 0.9. In cellular automata model, p_decAnd p_staTaken as 0.04 and 0.5, respectively.

TABLE 1 Nissan Leaf electric vehicle energy consumption model parameters

Parameter(s)	Numerical value	Parameter(s)	Numerical value
				A₁(N)	150	n_g	7.937
A₂(N/ms^-1)	0.61	η_g(％)	97
				A₃(N/m²s^-2)	0.51	η_m(％)	96
M(kg)	1521	PF_m	0.9
				R(m)	0.315	η_inv(％)	98
T_CF(N·m)	1.6	P_aux(kW)	0.18

By adopting the modeling method for the charging load of the quick charging stations, the charging load of each quick charging station in the area obtained by random optimization is shown in figures 6-9. A relatively small peak charge occurs around 6 am, which is mainly caused by the charging demand of class 1 taxis in part during the midnight. In addition, the peak value of the charging load in one day is mainly concentrated at 11-12 pm and 17-18 pm. The peak value at 11-12 noon is mainly caused by the charging demand of 2 types of taxis and the secondary charging demand of 1 type of taxis going out in the morning, 17-18 points are the peak period of traffic going out all day long, the energy consumption of the electric automobile going out is increased due to increased road section flow, and the charging demands of the electric automobile and the taxis are overlapped to generate a charging peak. The fast charge load gradually decreased after 20 pm, which is caused by the choice of slow charge by most electric vehicle users back home at night. Meanwhile, as can be seen from the simulation result, in the time period with a high fast-charging load level, the range of the uncertain interval is relatively large, and the corresponding relation between the strong charging demand and the charging place selection randomness is reflected.

The random/robust hybrid optimization scheduling effect is illustrated by taking the No. 1 power distribution network and the fast charging stations of the accessed

traffic nodes

6 and 7 as examples. The energy storage devices, SVCs and photovoltaic parameters accessed in the two fast charging stations are shown in Table 2.

TABLE 2 photovoltaic, energy storage and SVC device parameters

The photovoltaic output and the fluctuation deviation of the load power of the power distribution network are respectively 15% and 10% of the predicted value, and the specific prediction conditions are shown in figures 10-11. As shown in fig. 12, the random optimization scheduling effect is illustrated, the feeder capacity of the distribution point network No. 1 is 3.5kW, and under the condition that the value of the uncertain regulation parameter is maximum, the net load in the distribution network of each scene is out of limit. By adopting the optimized scheduling method provided by the invention, the energy storage charging and discharging power condition is shown as a bar chart in fig. 12, and it can be seen that the energy storage is mainly charged at the valley time of night and discharged at the moment of exceeding the limit, so that the capacity requirement in the power distribution network is met with lower cost, and the net load curve of the power distribution network under all scenes is smoother. To illustrate the robust optimization scheduling effect, three groups of 'uncertain adjustment parameters' are taken, and scheduling schemes are respectively compared: selecting gamma ^PV0 and Γ^LWhen the load is equal to 0, namely under the scene, the photovoltaic output and the original load of the power distribution network are predicted values; selecting gamma ^PV6 and Γ ^L12, the worst scenario obtained by optimization at this time is: the photovoltaic output is taken as the minimum value of a prediction interval in 6 time periods of 10-15h, and the original load of the power distribution network is taken as the maximum value of the prediction interval in 12 time periods of 9-13h and 17-23 h; selecting gamma ^PV24 and Γ ^L24, namely in the scene, boundary values are selected for the photovoltaic output and the original load of the power distribution network at each moment, and the scheme is the 'worst' scheme in all robust optimization schemes. Table 3 shows the different costs of the distribution network under different scenarios. It can be easily found that as the uncertainty adjusting parameter increases, the network loss cost, the electricity purchasing cost, the energy storage operation cost and the total operation cost of the power distribution network correspondingly increase. This is because as the optimal scheduling takes more into account the uncertainty faced by the system,the more conservative the obtained scheme is, the more the power purchasing amount and the network loss of the power distribution network to the upper-level power grid are increased, and the corresponding cost is also increased.

TABLE 3 comparison of costs under different protocols

The above-mentioned embodiments are only for illustrating the technical ideas and features of the present invention, and the purpose thereof is to enable those skilled in the art to understand the contents of the present invention and to carry out the same, and the present invention should not be limited to the embodiments, i.e. equivalent variations or modifications made within the spirit of the present invention are still within the scope of the present invention.

Claims

1. A power distribution network hybrid optimization scheduling method considering a light storage and fast charging integrated station is characterized by establishing a road traffic network model, a driving speed and path model of an electric automobile and an electric automobile travel time model, establishing a charging load model of the fast charging integrated station based on the three models, and processing uncertainty of fast charging load by adopting a random optimization method; on the basis of a charging load model of the quick charging integrated station, simultaneously adopting an uncertain set to represent photovoltaic output and the range of original load in the power distribution network, setting a constraint condition to establish a robust optimization scheduling model of the power distribution network of the quick charging integrated station by taking the minimum overall operation cost of the power distribution network to which the quick charging integrated station belongs as a target function; decomposing a power distribution network robust optimization scheduling model of the fast charging integration station into a main problem model and a sub problem model; the worst scene of the power distribution network tide operation is obtained by optimizing the photovoltaic output and the total time interval of the boundary value of the original load selection interval in the scheduling period; and solving the optimal solution of the main problem model and the sub problem model under the condition of meeting the convergence condition by adopting a column constraint generation algorithm to obtain a group of optimized scheduling strategy parameters consisting of main problem parameters and sub problem parameters.

2. The hybrid optimal scheduling method for the power distribution network considering the light storage and fast charging all-in-one station as claimed in claim 1, wherein a log-normal distribution and a normal distribution probability model are adopted to approximately simulate the corresponding initial trip time and return trip time distribution of the electric vehicle, and an electric vehicle trip time model is established as shown in the following formula:

3. The hybrid optimization scheduling method for the power distribution network considering the light storage and fast charging integrated station as claimed in claim 1, wherein the road traffic network model is established based on a classical graph theory method, and the relationship between traffic nodes and road segments is described by using an adjacency matrix D representing road weight, and the road traffic network model is established as shown in the following formula:

4. The hybrid optimization scheduling method for the power distribution network considering the light storage and fast charging all-in-one station as claimed in claim 1, wherein a driving speed and path model of the electric vehicle is established based on the shortest overall time principle.

5. The hybrid optimization scheduling method for the power distribution network considering the light storage and fast charging integrated station as claimed in claim 4, wherein the method for establishing the driving speed and path model of the electric vehicle is as follows:

in the formula, v_b(t) is the average speed of all electric vehicles in the section b at the moment t; v. of_i(t) is the speed of the ith vehicle at time t; n is a radical of_bRepresenting the number of electric vehicles on the road section b at the moment t; s_b(t) is the average transit time of the section b at the time t; l is_bRepresents the length of the section b;

in the formula, SOC_dRepresenting the state of charge of the electric vehicle when the destination is predicted to be reached; SOC_oThe state of charge of the electric vehicle as a starting point; s_dRepresents a distance traveled; c_EVRepresenting the battery capacity of the electric automobile; e represents the predicted value of the energy consumption of the electric automobile in unit kilometer;

if the remaining power is lower than the charging threshold value on the way, a proper charging station needs to be planned for charging, and at the moment, the queuing time and the charging time of the user in the charging station need to be considered, so that the driving speed and the path model of the electric vehicle are as follows:

indicating the queuing time at the j-th public charging station;

represents the charging time at the j-th public charging station;

6. The hybrid optimal scheduling method for the power distribution network considering the light storage and fast charging integrated station as claimed in claim 5, wherein a cellular automaton model is used to obtain the average speed of electric vehicles passing through a road section, and the specific method is as follows:

v_n(t+1)＝v_n(t)；

v_n(t+1)＝min{v_n(t)+1,v_max}；

v_n(t+1)＝min{v_n(t+1),d_n}；

v_n(t+1)＝max{v_n(t)-1,0}；

5) the position of the vehicle is updated as follows:

x_n(t+1)＝x_n(t)+v_n(t)。

7. the hybrid optimal scheduling method for the power distribution network considering the light storage and fast charging integrated station as claimed in claim 1, wherein an objective function of a robust optimal scheduling model for the power distribution network of the fast charging integrated station is as follows:

in the formula:

C_gridrepresenting the power purchase cost of the power distribution network;

C_lossrepresenting line loss cost;

C_storagerepresenting the cost of energy storage charging and discharging;

Ω_ta set of scheduling moments for one operating cycle of a typical scene;

Ω_essis an energy storage system set;

eta is the charge-discharge efficiency of the energy storage converter;

Ω_srepresents a collection of all scenes;

ω_sis the probability of occurrence of scene s;

the time-of-use electricity price at the time t is obtained;

K_lossis the unit loss cost;

R_ijrepresenting the resistance value of branch ij.

8. The method for hybrid optimal scheduling of the power distribution network in consideration of the light storage and fast charging integrated station according to claim 1, wherein the set constraint conditions include: the method comprises the following steps of power distribution network power flow operation restraint, energy storage device restraint and static reactive power compensation device restraint.

9. The hybrid optimal scheduling method for the power distribution network considering the light storage and fast charging integrated station as claimed in claim 8, wherein the power distribution network load flow operation constraint comprises:

u_min≤u_i,s,t≤u_max (s-7)；

0≤i_ij,s,t≤i_max (s-8)；

in the above formula:

i → j indicates that node i is the upstream node of node j;

j → k indicates that node j is an upstream node of node k;

i_maxa square value representing an upper limit of the branch current;

u_minis the square value of the lower limit of the node voltage;

u_maxis the square of the upper limit of the node voltage;

P_j,s,tthe net load active power of a node j at the moment t of a scene s;

Q_j,s,tthe net load reactive power of a node j at the moment t of a scene s;

R_ijthe resistance value of the line between the node i and the node j;

the active power of the original load of the node j at the moment t;

the reactive power of the original load of the node j at the moment t;

fast charging station charging load at a node j at the moment t of a scene s;

charging power of an energy storage system for a node j at the moment t;

the discharge power of the energy storage system is the node j at the moment t;

the active power of the photovoltaic output of the node j at the moment t.

10. The hybrid optimal scheduling method for the power distribution network considering the light storage and fast charging integrated station as claimed in claim 8, wherein the energy storage device constraint comprises:

in the above formula:

charging power of an energy storage system for a node j at the moment t;

the discharge power of the energy storage system is the node j at the moment t;

t represents the total number of scheduling periods;

storing the residual capacity of the node j at the initial scheduling moment;

represents the energy storage capacity of node j;

eta is the charge-discharge efficiency of the energy storage converter;

Δ t represents the simulation step size.

11. The power distribution network hybrid optimization scheduling method considering the light storage and fast charging integrated station is characterized in that a Monte Carlo simulation method is adopted to generate a charging load scene of a fast charging station in the power distribution network, a scene clustering method is adopted to obtain a typical scene, and the worst scene of power distribution network load flow operation is further obtained.

12. The hybrid optimal scheduling method for the power distribution network considering the light storage and fast charging integrated station as claimed in claim 1, wherein the following uncertain sets are adopted to represent the ranges of the photovoltaic output and the original load in the power distribution network:

in the formula (I), the compound is shown in the specification,

the active power of the original load of the node j at the time t,

the active power of photovoltaic output of a node j at the time t;

and

and

correspondingly representing the fluctuation deviation of the photovoltaic output and the original load power; gamma-shaped^PVThe numerical control method comprises the steps of (1) aiming at uncertain adjustment parameters introduced by photovoltaic output, wherein the numerical range is an integer from 0 to T, and the total number of time periods of the photovoltaic output taking the minimum value or the maximum value of a fluctuation interval in a scheduling period is represented; gamma-shaped^LThe method comprises the steps of introducing uncertain adjustment parameters aiming at original load power, wherein the value range of the uncertain adjustment parameters is an integer from 0 to T, and the total number of time periods of the original load power which takes the minimum value or the maximum value of a fluctuation interval in a scheduling period is represented; t represents the total number of scheduling periods; gamma-shaped^PVAnd Γ^LThe conservative property for adjusting the optimal solution is that the scheme obtained by the method is more conservative when the value is larger, and conversely, the scheme is more risky.

13. The hybrid optimization scheduling method for the power distribution network considering the light storage fast charging integrated station as claimed in claim 1, wherein a method for solving the optimal solution of the main problem model and the sub problem model under the condition of meeting the convergence condition by adopting a column constraint generation algorithm comprises the following steps:

step I: setting x as a parameter to be optimized of the main problem model; setting f (x) as an optimal value expression of the sub-problem model, and setting y and v as parameters to be optimized of the sub-problem model, wherein v represents the worst scene; to power distribution network whole that quick-charging integrated station belongs toThe operation cost of the body is set as LB ═ infinity at the lower boundary and UB ═ infinity at the upper boundary; setting the iteration number as l, and setting the initial value of l as 1; given an initial worst scenario value v₁；

Step II: setting the value of the worst scene as v in the first iteration_l: according to the worst scene v_lSolving the main problem model to obtain the optimal solution of x

The optimized value of the first iteration of the x parameter is obtained;

Step III: will be provided with

And worst case scenario

Order to

Returning to the step II;

step VI: and finishing the iteration.