CN106451443A - Electric power system recovering method for centralized electric vehicle charging station - Google Patents

Abstract

The invention provides an electric power system recovering method for a centralized electric vehicle charging station, and aims to solve the problem that a regional electric vehicle system without a regular black-start unit or insufficient in black-start unit capacity cannot recover normally. The electric power system recovering method includes that firstly, an available battery capacity of the charging station is modeled to obtain a starting capacity which can be provided for the charging station at a power cut moment of an electric power system; secondly, a net rack reconstruction model based on double-layer optimization is established, in an upper-layer model, a maximized generating capacity available for the electric power system is used as a target to determine a recovering time for a generator unit, and in a lower-layer model, the minimal sum of charging capacities of circuits is used as a target to determine a recovering path; thirdly, relevant uncertain factors are processed by chance constrain programming, a double-layer net rack reconstruction model optimization model based on chance constrain programming is established by means combining chance constrain programming with the double-layer optimization model, and then an optimization problem is solved by a modified particle swarm optimization method.

Description

A kind of meter and the power system recovery method of concentrated electric automobile charging station

Technical field

The invention belongs to power system recovery field, particularly to the electric power of a kind of meter and concentrated electric automobile charging station System recovery method.

Background technology

How quick large-scale blackout is after occurring, safely and steadly recovers power system is a very important problem. Power system recovery after having a power failure on a large scale generally can be divided into three Main Stage:Black starting-up stage, rack reconstruction stage and load are extensive The multiple stage.Be directed to both at home and abroad power system recovery method be all using conventional black starting-up unit be have a power failure on a large scale after system non-black Starting unit provides startup power.When not having in regional power system, black starting-up power supply or black starting-up unit capacity are not abundant When, the unit of this regional power system cannot provide enough startup power, cannot realize power system recovery.

Content of the invention

The present invention is directed to the problems referred to above it is proposed that the power system recovery side of a kind of meter and concentrated electric automobile charging station Method.

The purpose of the present invention is realized by following technological means：A kind of meter and the electricity of concentrated electric automobile charging station Force system restoration methods, comprise the following steps：

(1) collection power system topology information, line capacitance, circuit recover required time, the startup power of generating set, The climbing rate of generating set, peak power output, the capacity of concentrated charging station, the capacity of power distribution station, battery dis-tribution model, list The capacity of block battery, and the trip rule of analog subscriber.Joined according to the capacity of concentrated charging station, the capacity of power distribution station, battery Send the trip rule of pattern, the capacity of monoblock battery, user, the probability obtaining the available battery capacity of concentrated charging station divides Cloth.

(2) generate M primary at random.

(3) it is directed to each particle, carry out Ms Monte-Carlo step, obtain the available battery capacity of concentrated charging station.

(4) it is directed to and samples each time, call bi-level optimal model, obtain the available generating capacity of power system.

Bi-level optimal model can be described as：

In formula:W_totalIn the time interval [0, N studied after occurring by power outage_T] in, the available generating of system is held Amount.WithDistinguish upper true value in its domain of definition for representative function f (x) and lower true value.P_igenWhen () is t t Carve the output of unit i, when moment t is less than the unit starting time, unit output is 0, when moment t opens more than unit The dynamic time, when output is less than peak power output, the output of unit is multiplied by moment t and machine for this electromotor climbing rate The difference of group starting time；The climbing time sum obtaining divided by climbing rate with peak power output more than starting time as moment t When, the output of unit is the peak power output of unit.P_istartT () is the startup work(that t starts required for unit i Rate, when moment t is less than starting time, the startup power of unit is 0；It is otherwise the startup power collecting.N_resIt is system The quantity of interior unit to be restored；u_iT () represents the recovery state of unit i, if unit i has been turned on before t or t, Then u_i(t)=1, otherwise u_i(t)=0.A is penalty coefficient.T_opt=[t_opt1,t_opt2,…,t_optZ,]^TFor Z dimensional vector, simultaneously It is also the solution of upper strata Optimized model, its either element t_optiRepresent the optimum Startup time of unit i；T_actOpen up for considering electrical network Flutter each unit actual recovery moment of information.U (x, y) is defined binary function, and as x >=y, its functional value is 1；Otherwise For 0.Formula 1 and 2 is upper layer model and the underlying model of bilayer model respectively.

(5) available generating capacity step 4 being obtained carries out chance constraint inspection, calculates the target function value of each particleIt is the maximum under the β of confidence interval for the available generating capacity of power system to be restored.

β is the confidence level of object function.P_r{ } represents the probability that in set, event is set up；

(6) according to target function valueCalculate the fitness of each particle.

(7) M primary position and speed are updated, thus obtaining new particle.According to step 3-6, calculate after updating The fitness of each particle.

(8) repeat step (7), until it reaches population reproductive order of generation Mc.

(9) using particle maximum for fitness as this optimization problem optimal solution, the corresponding upper and lower layer model of this particle Solution corresponds respectively to unit starting moment and the restoration path of optimum.

(10) the unit starting moment of the optimum being obtained according to step 9 and restoration path, recover to power system.

The invention has the beneficial effects as follows, it is no black starting-up power supply or black starting-up unit using concentrated electric automobile charging station The not abundant regional power system of capacity provides unit starting power, and auxiliary power system is recovered.Concrete feature is as follows：

A models to charging station available battery capacity, obtains for the given system power failure available startup of moment charging station Capacity.The computational methods being proposed have taken into full account the trip rule of user and the dispensing model of battery, more fit actual.

B establishes the model for Power System Restoration based on dual-layer optimization, and in upper layer model, to maximize, system is available to be sent out For target, capacitance determines that generating set recovers the moment；In underlying model, true with the minimum target of line charging electric capacity sum Determine restoration path.Avoid and recover to recover the disadvantage of separately optimizing with circuit by unit in existing power system recovery method End.

C adopts chance constrained programming to process related uncertain factor, and the structure base that combines with bi-level optimal model In the double layer grid reconstruction and optimization model of chance constrained programming, and then this optimization problem is solved using modified particle swarm optiziation. It is introduced into chance constraint method and process the uncertain factor in model so that the power system recovery scheme finally giving meets necessarily Confidence interval, meet the requirement of power scheduling personnel, therefore methods herein is more practical.

Accompanying drawing content

Fig. 1 starts the time response of unit to be restored for electric automobile charging station；

Fig. 2 is the Optimizing Flow of the power system recovery method of meter and electric automobile charging station；

Fig. 3 New England 10 machine 39 node system；

Fig. 4 is the final optimum restoration path of generating node.

Specific embodiment

The present invention proposes a kind of meter and electric automobile concentrated charging station serves as the rack reconstruction and optimization of black starting-up power supply Method.First, establish the mathematical model of assessment electric automobile charging station available battery capacity.Then it is proposed that rack reconstructs Bi-level optimal model, its at the middle and upper levels model be used for optimizing the recovery order of generating node, lower floor is then used for determining restoration path. On this basis, for large-scale blackout occur the moment, charging station can with number of batteries, battery charge state etc. uncertain because Element, constructs the double layer grid reconstruction and optimization model based on chance constrained programming, to obtain the system meeting certain confidence level Restoration methods.Afterwards, solve proposed Optimized model using the method for modified particle swarm optiziation.The method includes following Step：

(1) collection power system topology information, line capacitance, circuit recover required time, the startup power of generating set, The climbing rate of generating set, peak power output, the capacity of concentrated charging station, the capacity of power distribution station, battery dis-tribution model, list The capacity of block battery, and the trip rule of analog subscriber.Joined according to the capacity of concentrated charging station, the capacity of power distribution station, battery Send the trip rule of pattern, the capacity of monoblock battery, user, the probability obtaining the available battery capacity of concentrated charging station divides Cloth.

Proposed method to be described taking New England's 10 machine 39 node system as a example.In figure 3 it is assumed that node 33 is collection Medium-sized electric automobile charging station place node；Node 30,31,32,34,35,36,37,38,39 is electromotor section to be restored Point.Line parameter circuit value and generator 's parameter are respectively as shown in Table 1 and Table 2.

The line parameter circuit value of table 1 New England 10 machine 39 node system

Table 2 generator 's parameter

(2) 20 primaries are generated at random.

(3) it is directed to each particle, carries out 1000 Monte-Carlo step, the available battery obtaining concentrated charging station holds Amount.

(4) it is directed to and samples each time, call bi-level optimal model, obtain the available generating capacity of power system.

Bi-level optimal model can be described as：

In formula:W_totalAfter occurring by power outage in time interval [0, the 180] min being studied, the available of system sends out Capacitance.WithDistinguish upper true value in its domain of definition for representative function f (x) and lower true value.P_igenT () is The output of t unit i, when moment t is less than the unit starting time, unit output is 0, when moment t is more than machine Group starting time, when output is less than peak power output, the output of unit is multiplied by moment t for this electromotor climbing rate Difference with the unit starting time；The climbing time obtaining divided by climbing rate with peak power output more than starting time as moment t During sum, the output of unit is the peak power output of unit.P_istartT () is that t starts opening required for unit i Dynamic power, when moment t is less than starting time, the startup power of unit is 0；It is otherwise the startup power collecting.N_resIt is The quantity of unit to be restored in system；u_iT () represents the recovery state of unit i, if unit i opens before t or t Dynamic, then u_i(t)=1, otherwise u_i(t)=0.A is penalty coefficient.T_opt=[t_opt1,t_opt2,…,t_optZ,]^TFor Z dimensional vector, with When be also upper strata Optimized model solution, its either element t_optiRepresent the optimum Startup time of unit i；T_actFor considering electrical network Each unit of topology information is actual to recover the moment.U (x, y) is defined binary function, and as x >=y, its functional value is 1；No It is then 0.Formula 1 and 2 is upper layer model and the underlying model of bilayer model respectively.

(5) available generating capacity step 4 being obtained carries out chance constraint inspection, calculates the target function value of each particleIt is the maximum under the β of confidence interval for the available generating capacity of power system to be restored.

β is the confidence level of object function, β=0.95.P_r{ } represents the probability that in set, event is set up；

(6) according to target function valueCalculate the fitness of each particle.

(7) 20 primary positions and speed are updated, thus obtaining new particle.According to step 3-6, calculate after updating The fitness of each particle.

(8) repeat step (7), until it reaches population reproductive order of generation 50.

(9) using particle maximum for fitness as this optimization problem optimal solution, the corresponding upper and lower layer model of this particle Solution corresponds respectively to unit starting moment and the restoration path of optimum, as shown in Table 3 and Figure 2.During according to optimum unit starting Carve and restoration path, power system is recovered.

Table 3 generating node recovers moment final optimization pass result

Generating node	Recover the moment (min)	The capacity (MW h) that charging station provides	Generating node	Recover the moment (min)	Charging station provides capacity (MW.h)
						34	8	14.1817	32	82	0
30	29	4.5067	38	92	0
						37	61	0	31	102	0
36	67	0	39	109	0
						35	75	0

Claims (1)

1. a kind of power system recovery method of meter and concentrated electric automobile charging station is it is characterised in that comprise the following steps：

(1) collection power system topology information, line capacitance, circuit recover required time, the startup power of generating set, generating The climbing rate of unit, peak power output, the capacity of concentrated charging station, the capacity of power distribution station, battery dis-tribution model, monolithic electricity The capacity in pond, and the trip rule of analog subscriber.According to the capacity of concentrated charging station, the capacity of power distribution station, battery dispensing mould Formula, the capacity of monoblock battery, the trip rule of user, obtain the probability distribution of the available battery capacity of concentrated charging station.

(2) generate M primary at random.

(3) it is directed to each particle, carry out Ms Monte-Carlo step, obtain the available battery capacity of concentrated charging station.

(4) it is directed to and samples each time, call bi-level optimal model, obtain the available generating capacity of power system.

Bi-level optimal model can be described as：

$W_{total}=\underset{T_{opt}}{sup}\{\underset{t=1}{\overset{N_T}{\Σ}}\underset{i=1}{\overset{N_{res}}{\Σ}}\[P_{igen}\left(t\right)-P_{istart}{u}_i\left(t\right)\]-aF\}---\left(1\right)$

$F=\underset{(T_{opt},T_{act})}{inf}\left\{\underset{i=1}{\overset{N_{res}}{\Σ}}U(t_{acti},t_{opti})\right\}---\left(2\right)$

In formula:W_totalIn the time interval [0, N studied after occurring by power outage_T] in, the available generating capacity of system.WithDistinguish upper true value in its domain of definition for representative function f (x) and lower true value.P_igenT () is t machine The output of group i, when moment t is less than the unit starting time, unit output is 0, when moment t is more than unit starting Between, when output is less than peak power output, the output of unit is multiplied by moment t for this electromotor climbing rate and is opened with unit The difference of dynamic time；When the climbing time sum that moment t is obtained divided by climbing rate with peak power output more than starting time, machine The output of group is the peak power output of unit.P_istartT () is the startup power that t starts required for unit i, when When moment t is less than starting time, the startup power of unit is 0；It is otherwise the startup power collecting.N_resIt is to treat in system Recover the quantity of unit；u_iT () represents the recovery state of unit i, if unit i has been turned on before t or t, u_i (t)=1, otherwise u_i(t)=0.A is penalty coefficient.T_opt=[t_opt1,t_opt2,…,t_optZ,]^TFor Z dimensional vector, it is also simultaneously The solution of upper strata Optimized model, its either element t_optiRepresent the optimum Startup time of unit i；T_actFor considering power network topology letter Each unit of breath is actual to recover the moment.U (x, y) is defined binary function, and as x >=y, its functional value is 1；It is otherwise 0. Formula 1 and 2 is upper layer model and the underlying model of bilayer model respectively.

(5) available generating capacity step 4 being obtained carries out chance constraint inspection, calculates the target function value of each particleI.e. Maximum under the β of confidence interval for the available generating capacity for power system to be restored.

$\begin{array}{cc}s.t.& P_r\{W_{total}\≥\stackrel{\‾}{W}\}\≥\mathrm{\β}\end{array}$

β is the confidence level of object function.P_r{ } represents the probability that in set, event is set up；

(6) according to target function valueCalculate the fitness of each particle.

(7) M primary position and speed are updated, thus obtaining new particle.According to step 3-6, calculate after updating each The fitness of particle.

(8) repeat step (7), until it reaches population reproductive order of generation Mc.

(9) using particle maximum for fitness as this optimization problem optimal solution, this particle corresponding levels solution to model divides Unit starting moment that Dui Yingyu be not optimum and restoration path.

(10) the unit starting moment of the optimum being obtained according to step 9 and restoration path, recover to power system.