CN104463701A - Coordinated planning method for power distribution system and electromobile charging network - Google Patents

Abstract

The invention discloses a coordinated planning method for a power distribution system and an electromobile charging network. The coordinated planning method comprises the following steps that first, a planning strategy of distributed charging piles based on node charging demands is put forward, and then a multi-objective optimization model of coordinated planning of the power distribution system and the electromobile charging network is established, wherein the objective of the model is that the sum of the investment cost and the system network loss is minimal, and the traffic flow intercepted and captured by quick charging stations is maximal. In the developed model, a traffic flow allocation model based on user balance is introduced to analyze the traffic flow intercepted and captured by the quick charging stations, and the number of devices of all the quick charging stations is allocated by the correlation theory of the queueing theory according to the rule that the average charging waiting time of vehicle owners is not beyond a certain threshold value at traffic rush hours. According to the method, a series of Pareto nondominated solutions can be obtained, planning personnel or investors can balance all the nondominated solutions according to practical engineering and select the most reasonable scheme, and therefore the investment efficiency is improved.

Description

A kind of coordinated planning method of distribution system and charging electric vehicle network

Technical field

The invention belongs to electric automobile charging station planning technology field, determine candidate site and the scale of quick charge station based on correlation theories such as the analysis of transportation network flow, waiting line theory, multiple-objection optimizations.

Background technology

The extensive access of electric automobile proposes new challenge to distribution system planning.First dispose charging electric vehicle network, and then to carry out upgrading to distribution system be a kind of two benches planing method comparatively conventional at present, but the method often can not realize the allocation optimum to resource effectively.Like this, how carrying out the coordinated planning of distribution system and charging electric vehicle network, is exactly a major issue being worth research.Under above-mentioned background, the present invention first proposed the planning strategy of the distributing charging pile based on node charge requirement, afterwards with cost of investment and system losses sum is minimum, quick charge station intercept and capture the magnitude of traffic flow be target to the maximum, construct the Model for Multi-Objective Optimization that distribution system and the charging electric vehicle network coordination are planned, and adopt the non-dominated sorted genetic algorithm-II (NSGA-II) of band amendment crossover and mutation operator to ask for the non-domination solution of constructed Model for Multi-Objective Optimization.

Summary of the invention

The distribution system of factor and the coordinated planning problems of charging electric vehicle network such as it is convenient that technical matters to be solved by this invention is the analysis of consideration urban traffic network flow, electric automobile user goes on a journey.

The coordinated planning method of distribution system of the present invention and charging electric vehicle network, comprises following 4 steps:

1) constant volume of distributing charging pile and the determination of charge power thereof

For distributing charging pile, the distributing charging pile quantity of the required configuration of each node of distribution system can be determined according to the charge requirement of each node.

$c_i=\mathrm{C\β}\frac{\underset{t\∈T}{\mathrm{\Σ}}{v}_{i,t}}{\underset{i\∈N^D}{\mathrm{\Σ}}\underset{t\∈T}{\mathrm{\Σ}}{v}_{i,t}},\∀i\∈N^D---\left(1\right)$

$n_i=\frac{c_i}{\mathrm{\γ}},\∀i\∈N^D---\left(2\right)$

In formula: c _iand n _icharging frequency demand (times/day) representing distribution system node i respectively and the distributing charging pile quantity that should configure; C represents total charging frequency demand of distribution system; β represents the ratio selecting distributing charging pile as energy recharge mode; v _i,trepresent that period t is in the electric automobile quantity of parked state in distribution system node i; N ^drepresent distribution system node set; T represents the period set in a day; γ represents the per day service times of single charging pile.

The distribution system introduced by distributing charging pile each node day part charge power and capacity thereof can be estimated according to formula (3)-(4) respectively.

${P_i}^{\mathrm{cp},\max }=n_i{p}^{\mathrm{cp}},\∀i\∈N^D---\left(4\right)$

In formula (3)-(4): with represent that the distribution system node i introduced by distributing charging pile is at the charge power of period t and capacity thereof respectively; H represents by the average charge duration of distributing charging pile as electric automobile energy supply mode; p ^cprepresent the specified charge power of distributing charging pile.

2) constant volume of quick charge station and the determination of charge power thereof

Assuming that the magnitude of traffic flow that the average arrival rate of each quick charge station vehicle to be charged and this quick charge station are intercepted and captured is directly proportional.Like this, the mean arrival rate (referring to arrive the electric automobile quantity that quick charge station accepts charging service in the unit interval) of quick charge station traffic peak period vehicle to be charged just can be expressed as:

${{\mathrm{\λ}}_k^{\mathrm{RH}}}_{}=\max \left\{{\mathrm{\λ}}_{k,t}\right|{\mathrm{\λ}}_{k,t}=C(1-\mathrm{\β})\frac{f_t^{\mathrm{trip}}}{\underset{t\∈T}{\mathrm{\Σ}}{f}_t^{\mathrm{trip}}}\frac{f_{k,t}}{\underset{k\∈{\mathrm{\Ω}}^K}{\mathrm{\Σ}}{f}_{k,t}},\∀t\∈T\},\∀k\∈{\mathrm{\Ω}}^K---\left(5\right)$

In formula: λ _k,twith represent that a kth rapid charge stands in the mean arrival rate (/ hour) of period t and traffic peak period vehicle to be charged respectively; Ω ^krepresent the candidate site set of quick charge station; f _t ^triprepresent the trip proportion of period t; f _k,trepresent the magnitude of traffic flow that a kth quick charge station period t intercepts and captures.

Certain threshold value is no more than for principle is to configure the number of devices of each quick charge station with the average charge stand-by period of traffic peak period car owner.According to the M/M/s queuing model in waiting line theory, (M/M/s represents that time of arrival is quantum condition entropy, service time is quantum condition entropy, service equipment quantity is the queuing system of s), the constant volume problem of quick charge station can be summed up as such as formula the nature of nonlinear integral programming problem shown in (6)-(10).

Objective function:

Min z _k(6)

Constraint condition:

$W_k^{\mathrm{RH}}<W^{\mathrm{allowed}},\&ForAll;k\&Element;{\mathrm{\&Omega;}}^K---\left(7\right)$

$W_k^{\mathrm{RH}}=\frac{{\left(z_k{\mathrm{\&rho;}}_k^{\mathrm{RH}}\right)}^{z_k}{\mathrm{\&rho;}}_k^{\mathrm{RH}}}{{\mathrm{\&lambda;}}_k^{\mathrm{RH}}\left(z_k\right)!{(1-{\mathrm{\&rho;}}_k^{\mathrm{RH}})}^2}{P}_{0,k},\&ForAll;k\&Element;{\mathrm{\&Omega;}}^K---\left(8\right)$

$P_{0,k}={[\underset{n=0}{\overset{z_k-1}{\mathrm{\&Sigma;}}}\frac{{\left(z_k{\mathrm{\&rho;}}_k^{\mathrm{RH}}\right)}^n}{n!}+\frac{{\left(z_k{\mathrm{\&rho;}}_k^{\mathrm{RH}}\right)}^{z_k}}{\left(z_k\right)!(1-{\mathrm{\&rho;}}_k^{\mathrm{RH}})}]}^{-1}---\left(9\right)$

${\mathrm{\&rho;}}_k^{\mathrm{RH}}=\frac{{\mathrm{\&lambda;}}_k^{\mathrm{RH}}}{z_k\mathrm{\&mu;}},\&ForAll;k\&Element;{\mathrm{\&Omega;}}^K---\left(10\right)$

In formula (6)-(10): z _krepresent the fast fill device quantity that a kth quick charge station should configure; and W ^allowedrepresent that the traffic peak period accepts the average latency of charging service and maximum permission time thereof respectively; represent the average service rate of the fast fill device of a traffic peak period kth quick charge station; P _{0, k}represent the probability that the fast fill device of a kth quick charge station is all idle; μ represents the mean service rate (/ hour) of fast fill device.

After determining the number of devices of each quick charge station, the charge power of its day part and maximum charge capacity can calculate according to formula (11)-(12) respectively.

$P_{k,t}^{\mathrm{FCS}}={\mathrm{\&rho;}}_{k,t}{z}_k{p}^{\mathrm{FCS}},\&ForAll;k\&Element;{\mathrm{\&Omega;}}^K,\&ForAll;t\&Element;T---\left(11\right)$

$P_k^{\mathrm{FCS},\max }=z_k{p}^{\mathrm{FCS}},\&ForAll;k\&Element;{\mathrm{\&Omega;}}^K---\left(12\right)$

In formula: with represent charge power and the maximum charge capacity thereof of a kth quick charge station period t respectively; p ^fCSrepresent the specified charge power of fast fill device; ρ _k,trepresent the equipment average service rate of a kth quick charge station period t, its available mathematical formulae is described as:

${\mathrm{\&rho;}}_{k,t}=\frac{{\mathrm{\&lambda;}}_{k,t}}{z_k\mathrm{\&mu;}},\&ForAll;k\&Element;{\mathrm{\&Omega;}}^K,\&ForAll;t\&Element;T---\left(13\right)$

3) Model for Multi-Objective Optimization that distribution system and the charging electric vehicle network coordination are planned is built

The target of first subproblem considering is for minimizing cost of investment and system losses sum, and its mathematical form can be described as:

Objective function:

$\begin{array}{c}\mathrm{Min}{f}_1={\mathrm{\&pi;}}^L\underset{\left(\mathrm{ij}\right)\&Element;{\mathrm{\&Omega;}}^{\mathrm{DL}}}{\mathrm{\&Sigma;}}{c}_{\mathrm{ij}}^L{x}_{\mathrm{ij}}{l}_{\mathrm{ij}}+{\mathrm{\&pi;}}^S(\underset{i\&Element;{\mathrm{\&Omega;}}^{\mathrm{SC}}}{\mathrm{\&Sigma;}}{c}_i^{\mathrm{SC}}{y}_i^{\mathrm{SC}}+\underset{i\&Element;{\mathrm{\&Omega;}}^{\mathrm{SR}}}{\mathrm{\&Sigma;}}{c}_i^{\mathrm{SR}}{y}_i^{\mathrm{SR}})\\ +{\mathrm{\&pi;}}^{\mathrm{FCS}}\underset{k\&Element;{\mathrm{\&Omega;}}^K}{\mathrm{\&Sigma;}}\left\{u_k\right[c^{\mathrm{CH}}{z}_k+c_k^{\mathrm{other}}{z}_k+c_k^F\left]\right\}\\ +c^E{d}^{\mathrm{annual}}\underset{\left(\mathrm{ij}\right)\&Element;{\mathrm{\&Omega;}}^{\mathrm{DL}}}{\mathrm{\&Sigma;}}\underset{t\&Element;T}{\mathrm{\&Sigma;}}[g_{\mathrm{ij}}{x}_{\mathrm{ij}}(U_{i,t}^2+U_{j,t}^2-2U_{i,t}{U}_{j,t}\cos {\mathrm{\&theta;}}_{\mathrm{ij},t})\end{array}---\left(14\right)$

Wherein:

${\mathrm{\&pi;}}^L=\frac{\mathrm{\&epsiv;}{(1+\mathrm{\&epsiv;})}^{n^L}}{{(1+\mathrm{\&epsiv;})}^{n^L}-1},{\mathrm{\&pi;}}^S=\frac{\mathrm{\&epsiv;}{(1+\mathrm{\&epsiv;})}^{n^S}}{{(1+\mathrm{\&epsiv;})}^{n^S}-1},{\mathrm{\&pi;}}^{\mathrm{FCS}}=\frac{\mathrm{\&epsiv;}{(1+\mathrm{\&epsiv;})}^{n^{\mathrm{FCS}}}}{{(1+\mathrm{\&epsiv;})}^{n^{\mathrm{FCS}}}-1}---\left(15\right)$

In formula: π ^l, π ^sand π ^fCSrepresent the recovery of the capital coefficient of distribution line, transformer station and quick charge station respectively; Ω ^dLrepresent distribution line set (envelope existing line and candidate line two class); Ω ^sRand Ω ^sErepresent respectively extendible capacity and not extendible capacity existing transformer station set; Ω ^sCrepresent the set of candidate transformer station; represent the construction cost (for existing line, this optimum configurations is 0) of distribution line; with represent newly-built cost and the dilatation cost of transformer station i respectively; c ^cHwith represent component (unit price as complete fast fill device) irrelevant with quick charge station geographic position in the variable cost that newly-built quick charge station relates to and the component (as land use cost) relevant with quick charge station geographic position respectively; represent the fixed cost (as auxiliary road construction cost and quick charge station inlet wire cost) of newly-built quick charge station; x _ij, and u _krepresent that distribution line construction, transformer station are newly-built respectively, binary decision variable that transformer station's dilatation and quick charge station are built; l _ijrepresent the length of distribution line ij; c ^eand d ^annualrepresent the average price of electric energy and the number of days in a year respectively; g _ijand b _ijrepresent conductance and the susceptance of distribution line ij respectively; G _ijand B _ijrepresent real part and the imaginary part of bus admittance matrix respectively; n ^l, n ^sand n ^fCSrepresent the length of service of distribution line, transformer station and quick charge station respectively; ε represents annual rate; U _i,tand U _j,trepresent the voltage magnitude of distribution system node i and node j period t respectively.

The target of second subproblem considering is the magnitude of traffic flow maximizing quick charge station intercepting and capturing, and its mathematical model can be described as:

Objective function:

$\begin{array}{cc}\mathrm{Max}& f_2=\underset{r\&Element;N^T}{\mathrm{\&Sigma;}}\underset{s\&Element;N^T}{\mathrm{\&Sigma;}}\underset{q\&Element;Q_{\mathrm{rs}}}{\mathrm{\&Sigma;}}{T}_{q,\mathrm{annual}}^{\mathrm{ts}}{\mathrm{\&tau;}}_q^{\mathrm{rs}}---\left(16\right)\end{array}$

Constraint condition:

$\underset{k\&Element;{\mathrm{\&Omega;}}_q^K}{\mathrm{\&Sigma;}}{u}_k\&GreaterEqual;{\mathrm{\&tau;}}_q^{\mathrm{rs}}---\left(17\right)$

$T_{q,\mathrm{annual}}^{\mathrm{rs}}=d^{\mathrm{annual}}\underset{t\&Element;T}{\mathrm{\&Sigma;}}{f}_{q,t}^{\mathrm{rs}}---\left(18\right)$

In formula: N ^trepresent the node set of transportation network; Q _rsrepresent the set of paths of starting point r and the terminal s (Origin Destination Pair rs, hereafter summary is that OD is to rs) connecting passenger car trip; represent the magnitude of traffic flow connecting the path q of OD to rs and carry every year; represent connect OD can by the decision variable intercepted and captured to the magnitude of traffic flow on the path q of rs, with represent that the magnitude of traffic flow on this path can and can not be intercepted and captured respectively; expression can intercept and capture the quick charge station set of the magnitude of traffic flow on the q of path; represent and connect OD to the magnitude of traffic flow of period t on the path q of rs.

4) quick non-dominated sorted genetic algorithm-II (Non-dominated sorting genetic algorithm II is adopted, NSGA-II) above-mentioned model is solved, a series of Pareto non-domination solution can be obtained, facilitate planning personnel or investor carry out weighing between each non-domination solution according to engineering reality and select the most reasonably scheme.

The invention has the beneficial effects as follows, numerical results shows, adopt the method proposed can obtain a series of Pareto non-domination solution, facilitate planning personnel or investor carry out weighing between each non-domination solution according to engineering reality and select the most reasonably scheme, thus improve efficiency of investment.

Accompanying drawing explanation

The coordinated planning method of Fig. 1 distribution system and charging electric vehicle network solve flow process

Fig. 2 is non-domination solution and the Pareto forward position of coordinated planning model

Fig. 3 is the topological structure of final plan

Fig. 4 is the voltage levvl of each node at virtual time slot

Fig. 5 is the applied power that each distribution line transmits at virtual time slot.

Embodiment

For the Construction Problems of electric automobile basis electrically-charging equipment, the present invention, on the basis of planning strategy providing the distributing charging pile based on node charge requirement, proposes with cost of investment and system losses sum is minimum, quick charge station is intercepted and captured the magnitude of traffic flow is the Model for Multi-Objective Optimization of target to the maximum.In constructed model, introduce the correlation theory of traffic flow distribution and waiting line theory, respectively the magnitude of traffic flow that each quick charge station is intercepted and captured is analyzed with the fast fill device quantity that should configure.Adopt the method proposed can obtain a series of Pareto non-domination solution, facilitate planning personnel or investor carry out weighing between each non-domination solution according to engineering reality and select the most reasonably scheme.

Adopt the coupled system of 54 Node power distribution system and 25 junction traffic networks so that the essential characteristic of the coordinated planning method of distribution system proposed by the invention and charging electric vehicle network to be described, here only using the candidate site of the common node of distribution system and transportation network as charging quickly station.The flow process that solves of the present invention is as shown in Fig. 1 in Figure of description, and physical planning result is as shown in Fig. 2-Fig. 5 in Figure of description.

Fig. 1: the coordinated planning method of distribution system and charging electric vehicle network solve in flow process, adopt the NSGA-II algorithm [20] of the crossover and mutation operator of band amendment to solve constructed Model for Multi-Objective Optimization.

Fig. 2: the Pareto forward position depicting non-domination solution, therefrom can find out, the increase of the magnitude of traffic flow intercepted and captured along with charging quickly station, cost of investment and system losses sum increase thereupon.Its reason mainly contains: the first, heavy traffic section, and the land use cost of its periphery is usually higher; The second, for whole distribution system, the programme being the charging quickly station of target with the magnitude of traffic flow intercepted and captured to the maximum often can not ensure minimizing of distribution system network loss.

Fig. 3: after given Pareto forward position, investor just can be actual according to engineering, weighs between each non-domination solution, selects the most reasonably scheme.For the coordinated planning problem of distribution system proposed by the invention and charging electric vehicle network, a kind of feasible choice criteria is: the minimum value (convenience degree of the service of filling soon provided to ensure electric automobile is not less than specified level) of the magnitude of traffic flow that given charging quickly station is intercepted and captured, and selects annual cost of investment and the minimum scheme of system losses sum.

Fig. 4 and Fig. 5: final plan is distinguished as shown in Figure 4 and Figure 5 at the voltage levvl of each node of virtual time slot and the applied power of each distribution line transmission.Therefrom known, in extreme circumstances, the program still can ensure the safe operation of distribution system.

Claims (1)

1. a coordinated planning method for distribution system and charging electric vehicle network, is characterized in that comprising following steps:

1) constant volume of distributing charging pile and the determination of charge power thereof

For distributing charging pile, the charging frequency demand c of distribution system node i every day _iand the distributing charging pile quantity n that should configure _ibe expressed as follows respectively:

In formula, C represents total charging frequency demand of distribution system; β represents the ratio selecting distributing charging pile as energy recharge mode; v _i,trepresent that period t is in the electric automobile quantity of parked state in distribution system node i; N ^drepresent distribution system node set; T represents the period set in a day; γ represents the per day service times of single charging pile;

The distribution system introduced by distributing charging pile each node day part charge power and capacity thereof can be estimated according to formula (3)-(4) respectively:

In formula (3)-(4), with represent that the distribution system node i introduced by distributing charging pile is at the charge power of period t and capacity thereof respectively; H represents by the average charge duration of distributing charging pile as electric automobile energy supply mode; p ^cprepresent the specified charge power of distributing charging pile;

2) constant volume of quick charge station and the determination of charge power thereof

Assuming that the magnitude of traffic flow that the average arrival rate of each quick charge station vehicle to be charged and this quick charge station are intercepted and captured is directly proportional, like this, the mean arrival rate of quick charge station traffic peak period vehicle to be charged, namely arriving the electric automobile quantity that quick charge station accepts charging service in the unit interval can be expressed as:

In formula: λ _k,twith represent that a kth rapid charge stands in the mean arrival rate of period t and traffic peak period vehicle to be charged per hour respectively; Ω ^krepresent the candidate site set of quick charge station; represent the trip proportion of period t; f _k,trepresent the magnitude of traffic flow that a kth quick charge station period t intercepts and captures;

According to the M/M/s queuing model in waiting line theory, the constant volume problem of quick charge station can be summed up as such as formula the nature of nonlinear integral programming problem shown in (6)-(10):

Min z _k(6)

In formula (6)-(10): z _krepresent the fast fill device quantity that a kth quick charge station should configure; and W ^allowedrepresent that the traffic peak period accepts the average latency of charging service and maximum permission time thereof respectively; represent the average service rate of the fast fill device of a traffic peak period kth quick charge station; P _{0, k}represent the probability that the fast fill device of a kth quick charge station is all idle; μ represents fast fill device mean service rate hourly;

After determining the number of devices of each quick charge station, the charge power of its day part and maximum charge capacity can calculate according to formula (11)-(12) respectively:

In formula: with represent charge power and the maximum charge capacity thereof of a kth quick charge station period t respectively; p ^fCSrepresent the specified charge power of fast fill device; ρ _{k, t}represent the equipment average service rate of a kth quick charge station period t, its available mathematical formulae is described as:

3) Model for Multi-Objective Optimization that distribution system and the charging electric vehicle network coordination are planned is built

The target of first subproblem considering is for minimizing cost of investment and system losses, and its mathematical form can be described as:

Wherein:

In formula: π ^l, π ^sand π ^fCSrepresent the recovery of the capital coefficient of distribution line, transformer station and quick charge station respectively; Ω ^dLrepresent the set of envelope existing line and candidate line two class distribution line; Ω ^sRand Ω ^sErepresent respectively extendible capacity and not extendible capacity existing transformer station set; Ω ^sCrepresent the set of candidate transformer station; represent the construction cost of distribution line, for existing line, this optimum configurations is 0; with represent newly-built cost and the dilatation cost of transformer station i respectively; c ^cHwith represent respectively in the variable cost that newly-built quick charge station relates to the component that quick charge station geographic position has nothing to do and the component relevant with quick charge station geographic position; represent the fixed cost of newly-built quick charge station; x _ij, and u _krepresent that distribution line construction, transformer station are newly-built respectively, binary decision variable that transformer station's dilatation and quick charge station are built; l _ijrepresent the length of distribution line ij; c ^eand d ^annualrepresent the average price of electric energy and the number of days in a year respectively; g _ijand b _ijrepresent conductance and the susceptance of distribution line ij respectively; G _ijand B _ijrepresent real part and the imaginary part of bus admittance matrix respectively; n ^l, n ^sand n ^fCSrepresent the length of service of distribution line, transformer station and quick charge station respectively; ε represents annual rate; U _i,tand U _j,trepresent the voltage magnitude of distribution system node i and node j period t respectively;

The target of second subproblem considering is the magnitude of traffic flow maximizing quick charge station intercepting and capturing, and its mathematical model can be described as:

In formula: N ^trepresent the node set of transportation network; Q _rsrepresent and connect OD to the set of paths of rs; represent the magnitude of traffic flow connecting the path q of OD to rs and carry every year; represent connect OD can by the decision variable intercepted and captured to the magnitude of traffic flow on the path q of rs, with represent that the magnitude of traffic flow on this path can and can not be intercepted and captured respectively; expression can intercept and capture the quick charge station set of the magnitude of traffic flow on the q of path; represent and connect OD to the magnitude of traffic flow of period t on the path q of rs;

4) adopt NSGA-II to solve above-mentioned model, obtain Pareto non-domination solution, planning personnel or investor weigh according to engineering reality and select the most reasonably scheme between each non-domination solution.