CN105389621A - Optimal charging pile distribution method for improving effect of electric vehicle charging load to voltage of distribution network system - Google Patents

Abstract

An optimal charging pile distribution method for improving an effect of an electric vehicle charging load to voltage of a distribution network system comprises the following steps of 1) calculating the node voltage of the distribution network system caused by the electric vehicle charging load at three conditions of different times, different load permeabilities and same-time same permeability; 2) according to a queuing theory, establishing a satisfaction function according to customer waiting time and charging pile utilization rate, and obtaining an optimal charging pile number; and 3) utilizing least node voltage drop as an objective function, and obtaining optimal distribution of the charging piles through simulating an annealing algorithm. Compared with an existing charging pile distribution method, the site distribution after optimization can greatly reduce the effect of the electric vehicle charging load to the node voltage of the distribution network system.

Description

Charging electric vehicle load is to the charging pile Optimal Distribution method of distribution network systems voltage influence

Technical field

The method that the site that the present invention relates to a kind of charging equipment of electric automobile is optimized, especially studies charging electric vehicle load to the method for the Optimal Distribution of the charging pile of distribution voltage influence based on simulated annealing.

Background technology

In July, 2015, Chinese Government proposed to meet at the year two thousand twenty the distribution network construction target that 1.2 ten thousand are filled electrical changing station, 4,800,000 charging pile access demands.But extensive electric automobile puts into operation and will produce many inevitable power quality problems to distribution.So, distribution network planning with fill change electric facilities planning synchronously seem particularly important, wherein the determination of electric automobile charging pile quantity and the planning and design of charging pile site, must consider the quality of voltage of public electric wire net.

Existing charging pile site optimize great majority be consider construction put into operation cost basis on carry out, as in " the city electric car charging station based on quantum particle swarm optimization is optimized distribution " paper that the people such as Liu Zifa delivered 2012 " Proceedings of the CSEE " the 32nd phase 39 pages to 45 pages, propose a kind of electric automobile charging station layout optimization method based on quanta particle swarm optimization, with cost of land, the construction costs such as the investment of substation transformer and comprise current supply loss operating cost based on set up objective function, do not consider that charging load is on the impact of distribution index, the grid structure of distribution is not considered yet.

In addition, determining charging pile quantitative aspects, also having different optimisation techniques.As the people such as Xiong Hu deliver in " optimal programming of electric automobile public charging station layout " paper of " Automation of Electric Systems " the 36th phase 65 pages to 70 pages for 2013, propose to consider that the quantity of charging pile is determined in the restriction of electric automobile queue time, but do not consider the utilization factor of charging station.

Summary of the invention

The object of the invention is for prior art not enough, provide a kind of charging electric vehicle load to the charging pile Optimal Distribution method of distribution network systems voltage influence.

The technical solution adopted for the present invention to solve the technical problems is: first not in the same time, under different load permeability and the same permeability of synchronization 3 kinds of sights, consider charging electric vehicle duration of load application and two, space dimension, analyze the impact in various degree that charging electric vehicle load causes distribution pressure drop, then according to waiting line theory, extent function is set up with Customer waiting time and charging pile utilization factor, obtain optimum charging pile quantity, and then it is minimum for objective function with node pressure drop, by simulated annealing, obtain the Optimal Distribution of charging pile.

Charging electric vehicle load of the present invention, to the charging pile Optimal Distribution method of distribution network systems voltage influence, comprises the following steps:

(1) calculate not in the same time, under different load permeability and the same permeability of synchronization 3 kinds of sights charging electric vehicle load to the pressure drop of distribution network systems node voltage;

First the charge model of electric automobile is set up.For distribution network systems, the charging electric vehicle load of access distribution, compared with traditional load, has randomness and ambulant feature, needs to emulate from Time and place two dimensions charging load, and then obtains the charging load of electric automobile every day.

Then under the optimization sight not having charging modes, each node of the charging load Stochastic accessing distribution network systems of electric automobile.Analyze different rush hour (8:00,12:00,18:00) permeability be under the sight of 30% charging electric vehicle load on the impact of system node voltage, under analyzing synchronization different load permeability (10%, 30%, 50%), charging electric vehicle load is on the impact of system node voltage, analyze the 8:00 moment, load permeability is 30%, the impact of charging electric vehicle load on system node voltage when concentration of local is distributed in individual node respectively, chooses and wherein analyzes the node that voltage influence degree is maximum, larger, minimum.Wherein during concentration of local distribution, assuming that 50% Assembled distribution of charging load is at supposition node, all the other are evenly distributed in all the other nodes.

The present invention adopts forward-backward sweep method computing node voltage electric current and voltage computing formula as follows:

$I_j^{*}=\frac{P_j^{*}-{\mathrm{jQ}}_j^{*}}{U_j^{*}}+\underset{m}{\overset{n}{\Σ}}{I}_m^{*}---\left(1\right)$

$U_j^{*}=U_i^{*}-I_j^{*}\×Z_{ij}^{*}---\left(2\right)$

In formula represent that node j injects the electric current perunit value sum of its next node layer, represent branch impedance perunit value sum between node i j.

Voltage converges criterion is:

$max\left(\right|U_n^{*}\left(k+1\right)-U_n^{*}\left(k\right)\left|\right)<\mathrm{\&epsiv;}---\left(3\right)$

In formula, k represents current iteration number of times.

(2) according to waiting line theory, set up extent function with Customer waiting time and charging pile utilization factor, obtain optimum charging pile quantity;

Under the prerequisite meeting client's charge requirement, the stand-by period of client be considered, in addition also will consider the utilization factor of charging pile.Utilization factor and the charging pile quantity of charging pile are inversely proportional to, and Customer waiting time is directly proportional to charging pile quantity.

Under the prerequisite of rational charging pile utilization factor, period of reservation of number is shorter, and customer satisfaction is higher.If customer satisfaction function M _ias follows:

M _i＝0.4×(W _i/W _max) ^-1+0.6×q _i(4)

Wherein W _i/ W _maxfor the standardization process of stand-by period, coefficient 0.4 and 0.6 is for save coefficient, and object is proportion of increasing operation rate.

(3) minimum for objective function with node pressure drop, by simulated annealing, obtain the Optimal Distribution of charging pile.

Charging pile Optimal Distribution model is minimum for target with the pressure drop of distribution network systems node, and objective function is:

$A=\min \underset{t=1}{\overset{t_{\max }}{\&Sigma;}}\underset{n=1}{\overset{N}{\&Sigma;}}{Z}_n{\mathrm{\&Delta;I}}_{n}t---\left(5\right)$

In formula: Z _nrepresent the impedance of distribution node n; Δ I _nrepresent the current differential that node n produces because of the access of charging load.

Simulated annealing is the intelligent algorithm searching globally optimal solution in solution space, enough large in initial temperature, and temperature declines under the most enough slow conditions, and energy convergence with probability 1 is to global optimum.

Beneficial effect of the present invention is as follows: 1) the present invention is when determining electric automobile charging pile, consider Time and place two dimensions of charging electric vehicle load, set up extent function according to electric automobile charging pile utilization factor and Customer waiting time, automatically provide the best configuration quantity of electric automobile charging pile; 2) the present invention when electric automobile accesses distribution network systems in a large number, when the load that charges affects minimum on the pressure drop of distribution network systems node, can provide the optimum site distribution of electric automobile charging pile automatically.

Accompanying drawing explanation

The algorithm flow schematic diagram of Fig. 1 the inventive method; Fig. 2 charging electric vehicle load curve;

Fig. 3 extent function operation result; Algorithm operation result when Fig. 4 configures 7 charging piles;

Algorithm operation result when Fig. 5 configures 10 charging piles.

Embodiment

Below in conjunction with accompanying drawing and embodiment, the invention will be further described.

The present invention for embodiment, obtains the charging pile site Optimal Distribution of IEEE33 node system with IEEE33 standard nodes system.

(1) step 1 calculate not in the same time, under different load permeability and the same permeability of synchronization 3 kinds of sights charging electric vehicle load to the pressure drop of distribution network systems node voltage;

First the charge model of electric automobile is set up.Obtain the parameter of Chinese dissimilar electric automobile according to table 1, and then obtain the charging load of electric automobile every day.

The Chinese dissimilar electric automobile setting parameter of table 1

The load curve of each node electric automobile every day can be obtained by following steps, as shown in Figure 2:

1) electric automobile total quantity in IEEE33 node system is obtained according to permeability ε then according to the accounting m of the known dissimilar electric automobile of prediction of " 2013 Chinese Auto Industry development report " _k, thus obtain the quantity n of the electric automobile of type k _k=m _k× n;

2) by minute in units of, charging load in the unit of account time wherein p _krepresent the charge power in the kth kind vehicle unit interval;

3) the charging load of every day is calculated obtain load curve.

Then under the optimization sight not having charging modes, each node of the charging load Stochastic accessing distribution network systems of electric automobile.Analyze different rush hour (8:00,12:00,18:00) permeability be under the sight of 30% charging electric vehicle load on the impact of system node voltage, under analyzing synchronization different load permeability (10%, 30%, 50%), charging electric vehicle load is on the impact of system node voltage, analyze the 8:00 moment, load permeability is 30%, the impact of charging electric vehicle load on system node voltage when concentration of local is distributed in individual node respectively, chooses and wherein analyzes the node that voltage influence degree is maximum, larger, minimum.In 3, under different sight, pressure drop is analyzed as follows.

Sight before grid-connected compared to electric automobile, in 3 charging peak periods, charging load to system node voltage influence less be 18:00, affecting larger is 12:00, and that have the greatest impact is 8:00, and its maximum pressure drop reaches 21.1%.

Under the sight of synchronization electric automobile Stochastic accessing distribution node, charging load permeability is higher, and larger on node voltage impact, load permeability is that under 10%, 30%, 50% sight, pressure drop maximal value is respectively 0.52%, 4.67%, 18.65%.

Minimum on the impact of distribution network systems node voltage when load concentration of local of charging is distributed in node 1, pressure drop maximal value is 2.49%, and node 1 is the power supply point of distribution network systems; Affect comparatively large when being distributed in node 17, pressure drop maximal value is 10.05%, and node 17 is intermediate nodes of distribution network systems; Have the greatest impact to distribution network systems node voltage when being distributed in node 32, pressure drop maximal value is 20.73%, and node 32 is positioned at the end of node system.

When the arbitrary node of charging electric vehicle load Stochastic accessing distribution, obvious pressure drop will be brought to the voltage of public electric wire net rush hour in charging, and to increase along with the increase of load permeability; Under concentration of local node charging sight, pressure drop influence degree is power supply node < intermediate node < endpoint node.

(2) step 2 is according to waiting line theory, sets up extent function with Customer waiting time and charging pile utilization factor, obtains optimum charging pile quantity;

Suppose that the number of units of charging pile is i, it is the Poisson distribution of w that electric automobile quantity obeys parameter, and the time that electric automobile accepts to charge is set to and obeys parameter is the negative exponential function of u.Electric automobile queuing model is Multiple server stations model (M/M/S), charging pile utilization factor q _ifor:

q _i＝w/(i×u)(6)

Client's average waiting time W _ifor:

$W_i=\frac{\&lsqb;\underset{k=0}{\overset{i-1}{\&Sigma;}}\frac{{(w/u)}^{^k}}{k!}\&times;{(w/u)}^{^i}+\frac{{(w/u)}^{^2i}}{i!\&times;(1-q_i)}\&rsqb;\&times;q_i}{w^2\&times;(1-q_i)}---\left(7\right)$

It should be noted that, the precondition that formula (1) and (2) are set up is q _i< 1

Suppose w=0.14, as shown in Figure 3, when charging pile quantity is less than or equal to 10, charging pile utilization factor reaches more than 70%; When charging pile quantity is increased to 10 from 7, Customer waiting time decreases 55.6%, and when being increased to 15 from 10, Customer waiting time only decreases 3.88%, and when configuration 10 charging piles, satisfaction is the highest.

(3) step 3 is minimum for objective function with node pressure drop, by simulated annealing, obtains the Optimal Distribution of charging pile.

Initial information comprises essential information and the algorithm initial value of electric automobile and IEEE33 node system.The essential information of electric automobile comprises the burden with power, SOC etc. of duration of charging, separate unit electric automobile; The essential information of node system comprises topological relation, line impedance value etc. between the node burden with power of power distribution network, node; Algorithm initial value comprises initial node set, initial temperature, final temperature.

First by front pushing back the Load flow calculation in generation, node pressure drop and the A of initial node set is obtained ₁, then produce random perturbation and obtain new node set, computing node pressure drop and A again ₂, gather with outstanding node and continue iteration, until meet convergence criterion T _k+1< T, obtains best node set.

According to operation result Fig. 4 and Fig. 5 and table 2 known, when charging pile configures 7, extent function value is 0.95, and target function value is 0.0278, charging pile arrange node serial number be 1.2.18.20.21.23.24; When charging pile configures 10, extent function value is 1.29, and target function value is 0.4061, and the node serial number that charging pile is arranged is 1.2.3.4.5.18.20.21.23.24.Two kinds of allocation plans are compared, and scheme two satisfaction is larger, and scheme one target function value is more outstanding; Scheme two is node 3,4,5 than scheme more than, is the node that distance power supply point is nearer.

Table 2 charging pile allocation plan and Analysis of Satisfaction

Claims (1)

1. charging electric vehicle load is to a charging pile Optimal Distribution method for distribution network systems voltage influence, it is characterized in that:

(1) calculate not in the same time, under different load permeability and the same permeability of synchronization 3 kinds of sights charging electric vehicle load to the pressure drop of distribution network systems node voltage;

First the charge model of electric automobile is set up; For distribution network systems, the charging electric vehicle load of access distribution, compared with traditional load, has randomness and ambulant feature, needs to emulate from Time and place two dimensions charging load, and then obtains the charging load of electric automobile every day;

Then under the optimization sight not having charging modes, each node of the charging load Stochastic accessing distribution network systems of electric automobile; Analyze different rush hour and 8:00,12:00,18:00 permeability be under the sight of 30% charging electric vehicle load on the impact of system node voltage, analyze synchronization different load permeability i.e. 10%, 30%, 50% time charging electric vehicle load to the impact of system node voltage, analyze the 8:00 moment, load permeability is 30%, the impact of charging electric vehicle load on system node voltage when concentration of local is distributed in individual node respectively, chooses and wherein analyzes the node that voltage influence degree is maximum, larger, minimum; Wherein during concentration of local distribution, assuming that 50% Assembled distribution of charging load is at supposition node, all the other are evenly distributed in all the other nodes;

The present invention adopts forward-backward sweep method computing node voltage electric current and voltage computing formula as follows:

$I_j^{*}=\frac{P_j^{*}-{\mathrm{jQ}}_j^{*}}{U_j^{*}}+\underset{m}{\overset{n}{\&Sigma;}}{I}_m^{*}---\left(1\right)$

$U_j^{*}=U_i^{*}-I_j^{*}\&times;Z_{ij}^{*}---\left(2\right)$

In formula represent that node j injects the electric current perunit value sum of its next node layer, represent branch impedance perunit value sum between node i j;

Voltage converges criterion is:

$max\left(\right|U_n^{*}(k+1)-U_n^{*}\left(k\right)\left|\right)<\mathrm{\&epsiv;}---\left(3\right)$

In formula, k represents current iteration number of times;

(2) according to waiting line theory, set up extent function with Customer waiting time and charging pile utilization factor, obtain optimum charging pile quantity;

Under the prerequisite meeting client's charge requirement, the stand-by period of client be considered, in addition also will consider the utilization factor of charging pile; Utilization factor and the charging pile quantity of charging pile are inversely proportional to, and Customer waiting time is directly proportional to charging pile quantity;

Under the prerequisite of rational charging pile utilization factor, period of reservation of number is shorter, and customer satisfaction is higher; If customer satisfaction function M _ias follows:

M _i＝0.4×(W _i/W _max) ^-1+0.6×q _i(4)

Wherein W _i/ W _maxfor the standardization process of stand-by period, coefficient 0.4 and 0.6 is for save coefficient, and object is proportion of increasing operation rate;

(3) minimum for objective function with node pressure drop, by simulated annealing, obtain the Optimal Distribution of charging pile;

Charging pile Optimal Distribution model is minimum for target with the pressure drop of distribution network systems node, and objective function is:

$A=min\underset{t=1}{\overset{t_{\max }}{\&Sigma;}}\underset{n=1}{\overset{N}{\&Sigma;}}{Z}_n{\mathrm{\&Delta;I}}_{n}t---\left(5\right)$

In formula: Z _nrepresent the impedance of distribution node n; Δ I _nrepresent the current differential that node n produces because of the access of charging load;

Simulated annealing is the intelligent algorithm searching globally optimal solution in solution space, enough large in initial temperature, and temperature declines under the most enough slow conditions, and energy convergence with probability 1 is to global optimum.