CN105405072A - Island heuristic load reduction model construction method for active power distribution network - Google Patents

Abstract

Disclosed is an island heuristic load reduction model construction method for an active power distribution network. The method comprises the following steps: constructing element models when an island runs, including a blower fan output model, a photovoltaic output model, a load model and a storage battery model; constructing a load reduction model; and solving an island load reduction model. According to the invention, the island heuristic load reduction model construction method for the active power distribution network is fast in operation speed and small in occupied memory, is used for rapidly determining a load reduction strategy, helps to improve the operation reliability of the active power distribution network, guarantees uninterrupted power supply of an important load, facilitates utilization of a renewable distributed power source, saves fossil energy and protects the ecological environment.

Description

The construction method of heuristic load summate model in the isolated island of active power distribution network

Technical field

The present invention relates to the load summate model that a kind of Power System Reliability is run.Particularly relate to a kind of for determining power distribution network islet operation time electric power supply not enough situation under the construction method of heuristic load summate model in the isolated island of active power distribution network of load summate strategy.

Background technology

In recent years, along with being on the rise of china natural resources environmental problem, the distributed power source being feature with resources conservation environmental friendliness obtains development energetically, has more and more accessed among existing power distribution network.The introducing of distributed power source improves the dirigibility of conventional electrical distribution net, economy and the feature of environmental protection, but also will change failure operation mode and the reliability estimation method of conventional electrical distribution net, need to consider the running statuses such as distributed power source grid-connected island, for the reliability service assessment of electric system brings new challenge.

Existing distributed power source is except typical blower fan, and outside the low capacity renewable energy power generation equipment such as photovoltaic, also arranging in pairs or groups with accumulator is the energy storage device of representative, to exert oneself the deficiency of undulatory property to make up regenerative resource.Accumulator, by adopting different discharge and recharge strategies, stores the excess power of distributed power source and load and discharges when undercapacity, maintains the even running of system under different conditions.

When a certain element failure of active power distribution network, in order to make full use of the features increase electric network reliability of distributed power source flexible operation, can fault occur after by the load of this locality with failure system electrical isolation, and utilizing local distributed power source separately for load is powered, local electric system is now in island operation state.When distributed power source is in liberal supply, islet operation can ensure that local load does not have a power failure by the impact of the system failure, after the system failure is repaired, the little electric system that distributed power source and local load form can be connected to the grid again.

Power-off condition in isolated island depends on inner power balance, and when distributed power source gross capability is greater than total load, the load point in isolated island can not dead electricity; And if during distributed power source undercapacity, need to carry out load summate, with balancing electric power supply and demand, the normal operation of guarantee section load.

Due to exerting oneself and load real-time change of distributed power source in isolated island, and accumulator is decided by the result of load summate in the peak power that each moment externally provides, and therefore load summate problem cannot be asked for by analytical method.And enumerative technique needs the longer time and calculates internal memory.Therefore, build a kind of Novel load towards active power distribution network and cut down model, to significantly improve arithmetic speed, there is good using value.

Summary of the invention

Technical matters to be solved by this invention is, provides one to have fast operation, the construction method of heuristic load summate model in the isolated island of active power distribution network that committed memory is little.

The technical solution adopted in the present invention is: a kind of construction method of heuristic load summate model in the isolated island of active power distribution network, comprises the steps:

1) component models when building islet operation, comprises exert oneself model, photovoltaic of blower fan and to exert oneself model, load model and battery model;

2) load summate model is built;

3) solve isolated island internal loading and cut down model.

Step 1) described in blower fan to exert oneself the structure of model, comprising:

(1) simulation of seasonal effect in time series arma modeling is adopted to produce the time series data of wind speed:

V _t＝μ _t+σ _ty _t(1)

y _t＝φ ₁y _t-1+φ ₂y _t-2+…φ _ly _t-l+…φ _ny _t-n+α _t-α _t-1θ ₁-α _t-2θ ₂-…α _t-sθ _s…-α _t-mθ _m(2)

In formula, V _tfor real-time wind speed; μ _tfor the mean value of historical wind speed data in assessment area, σ _tfor the standard deviation of historical wind speed distribution, y _tfor time series, φ _lfor autoregressive coefficient, l=1 ... n; θ _sfor running mean coefficient, s=1 ... m; α _tfor white noise coefficient, obedience average is 0, variance is independent normal distribution;

(2) set up blower fan to exert oneself model

$P_w=\{\begin{array}{cc}0,& 0\&le;V_t<V_{ci}\\ (A+B\&times;V_t+C\&times;{V_t}^2)P_r,& V_{ci}\&le;V_t<V_r\\ P_r,& V_r\&le;V_t\&le;V_{co}\\ 0,& V_t>V_{co}\end{array}---\left(3\right)$

In formula, P _wfor exerting oneself in real time of blower fan, A, B, C are the coefficient of the fitting function of power curve non-linear partial, V _tbe the real-time air speed data of t hour, V _ci, V _rand V _cobe respectively the incision wind speed of blower fan, wind rating and cut-out wind speed, P _rfor the output rating of blower fan.

Step 1) described in photovoltaic model of exerting oneself be:

$P_b=\left\{\begin{array}{cc}{P}_{sn}(G_{bt}^2/(G_{std}{R}_c\left)\right),& 0\&le;G_{bt}<R_c\\ P_{sn}(G_{bt}/G_{std}),& R_c\&le;G_{bt}<G_{std}\\ P_{sn},& G_{bt}\&GreaterEqual;G_{std}\end{array}\right.---\left(4\right)$

In formula, P _bfor exerting oneself in real time of photovoltaic, unit kW; P _snfor the rated power of photovoltaic, represent the power that unit light intensity can produce under standard test condition; G _stdfor specified intensity of illumination, unit is kW/m ²; R _cfor the light intensity of a certain certain strength, under this light intensity photovoltaic exert oneself to start with the relation of light intensity from non-linear become linear; G _btbe the real-time light intensity of t hour, unit kW/m ², G _btreal time sequence obtained by the sampling of the probability distribution statistical to history light intensity.

Step 1) described in load model be logical overladen Typical Year-all curves, week-curve and day-hour curve forms real-time load data and obtains:

L _t＝L _p×P _w×P _d×P _h(t)(5)

In formula, L _pfor year load peak, P _wfor corresponding with t hour year-all load curves in value, P _dfor corresponding with t hour week-daily load curve in value, P _h(t) be corresponding with t hour day-hour load curve in value.

Step 1) described in battery model be adopt two pool models of lead-acid accumulator, namely accumulator is divided into the ponds of two series connection mutually, and be wherein utilisable energy the first pond in, expression can be converted into the energy of electric energy immediately; Be bound energy in second pond, represent and to be subject in pond chemical reaction velocity constraint and electric energy cannot be converted at once or be the energy of chemical energy by electric energy conversion;

Battery discharging model representation is:

P _dmax＝min(P _dkbm,P _misoc)η _d(6)

In formula, P _dmaxrepresent the maximum continuous discharge power that accumulator is external, P _dkbmrepresent the battery discharging restrain condition described by two pool models of accumulator, P _misocrepresent minimum capacity constraint, i.e. the minimum state-of-charge of accumulator, η _dfor discharging efficiency, P _dkbmand P _misoccomputing formula as follows:

$P_{dkbm}=\frac{{\mathrm{kQ}}_1{e}^{-k\mathrm{\&Delta;}t}+Qkc(1-e^{-k\mathrm{\&Delta;}t})}{1-e^{-k\mathrm{\&Delta;}t}+c(k\mathrm{\&Delta;}t-1+e^{-k\mathrm{\&Delta;}t})}---\left(7\right)$

P _misoc＝Q _max(Soc _st-Soc _min)/Δt(8)

Wherein Q is the total surplus capacity of accumulator before electric discharge, Q ₁for the residual capacity in front first pond of discharging, Soc _minfor the minimum state-of-charge constraint of accumulator, Soc _stfor the state-of-charge of the front accumulator that charges, Δ t is any time period, Q _maxbe the total volume in two ponds, i.e. the rated capacity (kWh) of accumulator; C is the ratio of the first capacity and total volume, represents the capacity ratio in two ponds; K is constant (1/h), in order to the conversion rate of Characterization Energy between Liang Chi.

Charge in batteries model representation is:

P _cmax＝max(P _ckbm,-P _mcr,-P _mcc,-P _masoc)/η _c(9)

In formula, P _cmaxfor the maximum acceptable lasting charge power that accumulator is external, P _ckbmrepresent the charge in batteries restrain condition described by two pool models, P _mcrrepresent the maximum charge rate constraint of accumulator, P _mccrepresent the maximum permission charging current constraint of accumulator, P _masocrepresent max cap. constraint, i.e. the most highly charged state of accumulator, η _cfor charge efficiency.The computing method of each constraint are as follows:

$P_{ckbm}=\frac{-{\mathrm{kcQ}}_{\max }+{\mathrm{kQ}}_1{e}^{-k\mathrm{\&Delta;}t}+Qkc(1-e^{-k\mathrm{\&Delta;}t})}{1-e^{-k\mathrm{\&Delta;}t}+c(k\mathrm{\&Delta;}t-1+e^{-k\mathrm{\&Delta;}t})}---\left(10\right)$

P _mcr＝(1-e ^-αΔt)(Q _max-Q)/Δt(11)

P _mcc＝I _maxV _nom/1000(12)

P _masoc＝Q _max(Soc _max-Soc _st)/Δt(13)

Wherein, I _maxfor the maximum permission charging current of accumulator, V _nomfor accumulator rated operational voltage, Soc _maxfor the most highly charged state constraint.

Step 2) described in structure load summate model, first the condition of load summate is carried out in setting: the total load be more than or equal in this moment island of exerting oneself of the peak power that during islet operation, in island, all battery pack can discharge and all distributed power sources, and mathematic(al) representation is as follows:

$\begin{array}{c}\underset{i=1}{\overset{N_b}{\&Sigma;}}{P}_{d\max }^i\left(t\right)+\underset{j=1}{\overset{N_{DG}}{\&Sigma;}}{P}_{DG}^j\left(t\right)\&GreaterEqual;S_l\left(t\right)+\underset{k=1}{\overset{N_L}{\&Sigma;}}{L}_t^k\left(t\right)X\left(k\right),t\&Element;\&lsqb;t_{st},t_{end}\&rsqb;\\ when{S}_l\left(t\right)+\underset{k=1}{\overset{N_L}{\&Sigma;}}{L}_t^k\left(t\right)X\left(k\right)-\underset{j=1}{\overset{N_{DG}}{\&Sigma;}}{P}_{DG}^j\left(t\right)\&GreaterEqual;0\end{array}---\left(14\right)$

In formula, for jth distributed power source exerting oneself in real time in t, blower fan by blower fan exert oneself model calculate, photovoltaic by photovoltaic exert oneself model calculate, N _dGfor the sum of distributed power source; for the Real-time Load of a kth load point, calculated by load model, N _lfor the sum of load point; for the peak power that t i-th battery pack can externally provide in time interval Δ t, calculated by battery model, N _bfor total number of battery pack in isolated island, S _lt () is the active loss in isolated island, t _stfor isolated island forms the moment, t _endfor islet operation finish time, the time interval, Δ t got 1 hour, and X (k) is the reduction state of a kth load point.

If do not meet the constraint of the condition of formula load summate, just need the load cutting down a part of load point, to guarantee the power balance in isolated island, the objective function of load summate is:

$max\underset{t=t_{st}}{\overset{t}{\&Sigma;}}\underset{k=1}{\overset{N_L}{\&Sigma;}}{L}_t^k\left(t\right)X\left(k\right)---\left(15\right)$

In formula, X (k) is the reduction state of the kth load point in the condition of load summate, and X (k)=0 represents load point k and cut down, and X (k)=1 represents load point k and is retained;

The condition of load summate and the objective function of load summate constitute the load summate model in isolated island, and the objective function of load summate is the objective function of model, and the condition of load summate is constraint condition.

Step 3) described in solve isolated island internal loading cut down model, comprise the steps:

(1) reduction state X (k) setting each load point is 1, and calculates the total electricity W of each load point during isolated island according to the following formula _k,

$W_k=\underset{t=t_{st}}{\overset{t_{end}}{\&Sigma;}}{L}_t^k\left(t\right)---\left(16\right)$

(2) time t=t is made _st;

(3) the clean exchange power P in t isolated island is calculated according to the following formula _ex(t),

$P_{ex}\left(t\right)=\underset{k=1}{\overset{N_L}{\&Sigma;}}{L}_t^k\left(t\right)X\left(k\right)+S_l\left(t\right)-\underset{j=1}{\overset{N_{DG}}{\&Sigma;}}{P}_{DG}^j\left(t\right)---\left(17\right)$

If P _ext () < 0, performs step (4), otherwise, perform step (5);

(4) first, according to the rated capacity Q of each battery pack in t isolated island _{max, i}, state-of-charge Soc _ithe residual capacity Q in (t) and the first pond _{1, i}t (), utilizes following formula to try to achieve each accumulator lasting charge power P externally maximum acceptable in time interval Δ t _cmax(t);

Secondly, according to P _cmaxt the size of (), utilizes following formula to calculate the clean exchange power corresponding with each battery pack

$P_{ex}^i\left(t\right)=\frac{P_{cmax}^i\left(t\right)\&times;Q_{max,i}}{\underset{i=1}{\overset{N_b}{\&Sigma;}}{P}_{cmax}^i\left(t\right)\&times;Q_{max,i}}---\left(18\right)$

In formula, Q _{max, i}for the rated capacity of battery pack,

Finally, following formula is utilized to calculate the power of each battery pack actual absorption in Δ t

$P=\left\{\begin{array}{cc}\max (P_{c\max },P_{ex}^t){\mathrm{\&eta;}}_c& P_{ex}^t<0\\ \min (P_{d\max },P_{ex}^t)/{\mathrm{\&eta;}}_c& P_{ex}^t\&GreaterEqual;0\end{array}\right.---\left(19\right)$

And upgrade the state-of-charge Soc in t+ Δ t of each battery pack _ithe residual capacity Q in (t+ Δ t) and the first pond _{1, i}(t+ Δ t);

(5) first, according to the rated capacity Q of each battery pack in t isolated island _{max, i}, state-of-charge Soc _ithe residual capacity Q in (t) and the first pond _{1, i}(t), the peak power utilizing battery model to try to achieve each accumulator can externally to provide in time interval Δ t now, if size meet the constraint of load summate condition, then perform step (6); Otherwise, perform step (7);

(6) first, according to size, utilize following formula to calculate the clean exchange power corresponding with each battery pack

$P_{ex}^i\left(t\right)=\frac{P_{d\max }^i\left(t\right)\&times;Q_{\max ,i}}{\underset{i=1}{\overset{N_b}{\&Sigma;}}{P}_{d\max }^i\left(t\right)\&times;Q_{\max ,i}}---\left(20\right)$

Secondly, formula is utilized $P=\left\{\begin{array}{cc}\max (P_{c\max },P_{ex}^t){\mathrm{\&eta;}}_c& P_{ex}^t<0\\ \min (P_{d\max },P_{ex}^t)/{\mathrm{\&eta;}}_c& P_{ex}^t\&GreaterEqual;0\end{array}\right.$ Calculate the power of the actual release of each battery pack in Δ t and upgrade the state-of-charge Soc in t+ Δ t of each battery pack _ithe residual capacity Q in (t+ Δ t) and the first pond _{1, i}(t+ Δ t); Finally, time t=t is made _st+ Δ t, and return step (3);

(7) by all load point be retained, total electricity W _kreduction state X (k) of minimum load point is set as 0, as time t < t _endtime, return step (2); As time t=t _endtime, load summate flow process terminates, and records reduction state X (k) of final each load point.

The construction method of heuristic load summate model in the isolated island of active power distribution network of the present invention; fast operation; committed memory is little; for determining load summate strategy fast; contribute to the operational reliability promoting active power distribution network, ensure the uninterrupted power supply of important load, promote the utilization of renewable distributed power source; thus saving fossil energy, preserve the ecological environment.

Accompanying drawing explanation

Fig. 1 is the process flow diagram of the present invention's construction method of heuristic load summate model in the isolated island of active power distribution network;

Fig. 2 is the RBTSBus6 feeder system structural representation improved;

Fig. 3 is the annual power off time comparison diagram of load point in 2 kinds of situations.

Embodiment

Below in conjunction with embodiment and accompanying drawing, the construction method to heuristic load summate model in the isolated island of active power distribution network of the present invention is described in detail.

The construction method of heuristic load summate model in the isolated island of active power distribution network of the present invention, comprises the steps:

1) component models when building islet operation, comprises exert oneself model, photovoltaic of blower fan and to exert oneself model, load model and battery model; When active power distribution network is in islet operation pattern, isolated island inner member comprises the renewable distributed power source such as blower fan, photovoltaic and accumulator, load etc., and building its component models is the prerequisite setting up load summate model.Wherein:

Described blower fan is exerted oneself the structure of model, comprising:

(1) simulation of seasonal effect in time series arma modeling is adopted to produce the time series data of wind speed:

V _t＝μ _t+σ _ty _t(1)

y _t＝φ ₁y _t-1+φ ₂y _t-2+…φ _ny _t-n+α _t-α _t-1θ ₁-α _t-2θ ₂-…-α _t-mθ _m(2)

In formula, V _tfor real-time wind speed; μ _tfor the mean value of historical wind speed data in assessment area, σ _tfor the standard deviation of historical wind speed distribution, y _tfor time series, φ _lfor autoregressive coefficient, l=1 ... n; θ _sfor running mean coefficient, s=1 ... m; α _tfor white noise coefficient, obedience average is 0, variance is independent normal distribution;

(2) set up blower fan to exert oneself model

$P_w=\{\begin{array}{cc}0,& 0\&le;V_t<V_{ci}\\ (A+B\&times;V_t+C\&times;{V_t}^2)P_r,& V_{ci}\&le;V_t<V_r\\ P_r,& V_r\&le;V_t\&le;V_{co}\\ 0,& V_t>V_{co}\end{array}---\left(3\right)$

In formula, P _wfor exerting oneself in real time of blower fan, A, B, C are the coefficient of the fitting function of power curve non-linear partial, V _tbe the real-time air speed data of t hour, V _ci, V _rand V _cobe respectively the incision wind speed of blower fan, wind rating and cut-out wind speed, P _rfor the output rating of blower fan.

Described photovoltaic model of exerting oneself is:

$P_b=\left\{\begin{array}{cc}{P}_{sn}(G_{bt}^2/(G_{std}{R}_c\left)\right),& 0\&le;G_{bt}<R_c\\ P_{sn}(G_{bt}/G_{std}),& R_c\&le;G_{bt}<G_{std}\\ P_{sn},& G_{bt}\&GreaterEqual;G_{std}\end{array}\right.---\left(4\right)$

In formula, P _bfor exerting oneself in real time of photovoltaic, unit kW; P _snfor the rated power of photovoltaic, represent the power that unit light intensity can produce under standard test condition; G _stdfor specified intensity of illumination, unit is kW/m ²; R _cfor the light intensity of a certain certain strength, under this light intensity photovoltaic exert oneself to start with the relation of light intensity from non-linear become linear; G _btbe the real-time light intensity of t hour, unit kW/m ², G _btreal time sequence obtained by the sampling of the probability distribution statistical to history light intensity.

Described load model is logical overladen Typical Year-all curves, week-curve and day-hour curve forms real-time load data and obtains:

L _t＝L _p×P _w×P _d×P _h(t)(5)

In formula, L _pfor year load peak, P _wfor corresponding with t hour year-all load curves in value, P _dfor corresponding with t hour week-daily load curve in value, P _h(t) be corresponding with t hour day-hour load curve in value.

Described battery model is the two pool model (KineticBatteryModel adopting lead-acid accumulator, KiBaM), namely accumulator is divided into two ponds of mutually connecting, be wherein utilisable energy (avaliableenergy) in the first pond, represent and can be converted into the energy of electric energy immediately; Be bound energy (boundenergy) in second pond, represent and to be subject in pond chemical reaction velocity constraint and electric energy cannot be converted at once or be the energy of chemical energy by electric energy conversion;

Battery discharging model representation is:

P _dmax＝min(P _dkbm,P _misoc)η _d(6)

In formula, P _dmaxrepresent the maximum continuous discharge power that accumulator is external, P _dkbmrepresent the battery discharging restrain condition described by two pool models of accumulator, P _misocrepresent minimum capacity constraint, i.e. the minimum state-of-charge of accumulator, η _dfor discharging efficiency, P _dkbmand P _misoccomputing formula as follows:

$P_{dkbm}=\frac{{\mathrm{kQ}}_1{e}^{-k\mathrm{\&Delta;}t}+Qkc(1-e^{-k\mathrm{\&Delta;}t})}{1-e^{-k\mathrm{\&Delta;}t}+c(k\mathrm{\&Delta;}t-1+e^{-k\mathrm{\&Delta;}t})}---\left(7\right)$

P _misoc＝Q _max(Soc _st-Soc _min)/Δt(8)

Wherein Q is the total surplus capacity of accumulator before electric discharge, Q ₁for the residual capacity in front first pond of discharging, Soc _minfor the minimum state-of-charge constraint of accumulator, Soc _stfor the state-of-charge of the front accumulator that charges, Δ t is any time period, Q _maxbe the total volume in two ponds, i.e. the rated capacity (kWh) of accumulator; C is the ratio of the first capacity and total volume, represents the capacity ratio in two ponds; K is constant (1/h), in order to the conversion rate of Characterization Energy between Liang Chi.

Charge in batteries model representation is:

P _cmax＝max(P _ckbm,-P _mcr,-P _mcc,-P _masoc)/η _c(9)

In formula, P _cmaxfor the maximum acceptable lasting charge power that accumulator is external, P _ckbmrepresent the charge in batteries restrain condition described by two pool models, P _mcrrepresent the maximum charge rate constraint of accumulator, P _mccrepresent the maximum permission charging current constraint of accumulator, P _masocrepresent max cap. constraint, i.e. the most highly charged state of accumulator, η _cfor charge efficiency.The computing method of each constraint are as follows:

$P_{ckbm}=\frac{-{\mathrm{kcQ}}_{\max }+{\mathrm{kQ}}_1{e}^{-k\mathrm{\&Delta;}t}+Qkc(1-e^{-k\mathrm{\&Delta;}t})}{1-e^{-k\mathrm{\&Delta;}t}+c(k\mathrm{\&Delta;}t-1+e^{-k\mathrm{\&Delta;}t})}---\left(10\right)$

P _mcr＝(1-e ^-αΔt)(Q _max-Q)/Δt(11)

P _mcc＝I _maxV _nom/1000(12)

P _masoc＝Q _max(Soc _max-Soc _st)/Δt(13)

Wherein, I _maxfor the maximum permission charging current of accumulator, V _nomfor accumulator rated operational voltage, Soc _maxfor the most highly charged state constraint.

2) load summate model is built;

First the condition of load summate is carried out in setting: the total load be more than or equal in this moment island of exerting oneself of the peak power that during islet operation, in island, all battery pack can discharge and all distributed power sources, and mathematic(al) representation is as follows:

$\begin{array}{c}\underset{i=1}{\overset{N_b}{\&Sigma;}}{P}_{d\max }^i\left(t\right)+\underset{j=1}{\overset{N_{DG}}{\&Sigma;}}{P}_{DG}^j\left(t\right)\&GreaterEqual;S_l\left(t\right)+\underset{k=1}{\overset{N_L}{\&Sigma;}}{L}_t^k\left(t\right)X\left(k\right),t\&Element;\&lsqb;t_{st},t_{end}\&rsqb;\\ when{S}_l\left(t\right)+\underset{k=1}{\overset{N_L}{\&Sigma;}}{L}_t^k\left(t\right)X\left(k\right)-\underset{j=1}{\overset{N_{DG}}{\&Sigma;}}{P}_{DG}^j\left(t\right)\&GreaterEqual;0\end{array}---\left(14\right)$

In formula, for jth distributed power source exerting oneself in real time in t, blower fan by blower fan exert oneself model calculate, photovoltaic by photovoltaic exert oneself model calculate, N _dGfor the sum of distributed power source; for the Real-time Load of a kth load point, calculated by load model, N _lfor the sum of load point; for the peak power that t i-th battery pack can externally provide in time interval Δ t, calculated by battery model, N _bfor total number of battery pack in isolated island, S _lt () is the active loss in isolated island, t _stfor isolated island forms the moment, t _endfor islet operation finish time, the time interval, Δ t got 1 hour, and X (k) is the reduction state of a kth load point.

If do not meet the constraint of the condition of formula load summate, just need the load cutting down a part of load point, to guarantee the power balance in isolated island, the objective function of load summate is:

$max\underset{t=t_{st}}{\overset{t}{\&Sigma;}}\underset{k=1}{\overset{N_L}{\&Sigma;}}{L}_t^k\left(t\right)X\left(k\right)---\left(15\right)$

In formula, X (k) is the reduction state of the kth load point in the condition of load summate, and X (k)=0 represents load point k and cut down, and X (k)=1 represents load point k and is retained;

The condition of load summate and the objective function of load summate constitute the load summate model in isolated island, and the objective function of load summate is the objective function of model, and the condition of load summate is constraint condition.

3) solve isolated island internal loading and cut down model,

Solve isolated island internal loading and cut down model, first suppose that reduction state X (k) of all load point is 1, for [t during isolated island _st, t _end] in each moment, adopt above-mentioned scene to store model and calculate exerting oneself of each distributed power source and the state-of-charge of battery pack, when occurring not meeting the constraint of formula (9), the load point that during preferential reduction isolated island, internal loading total amount is minimum also returns t _stin the moment, repeat this process until meet formula (9).Specifically comprise the steps:

(1) reduction state X (k) setting each load point is 1, and calculates the total electricity W of each load point during isolated island according to the following formula _k,

$W_k=\underset{t=t_{st}}{\overset{t_{end}}{\&Sigma;}}{L}_t^k\left(t\right)---\left(16\right)$

(2) time t=t is made _st;

(3) the clean exchange power P in t isolated island is calculated according to the following formula _ex(t),

$P_{ex}\left(t\right)=\underset{k=1}{\overset{N_L}{\&Sigma;}}{L}_t^k\left(t\right)X\left(k\right)+S_l\left(t\right)-\underset{j=1}{\overset{N_{DG}}{\&Sigma;}}{P}_{DG}^j\left(t\right)---\left(17\right)$

If P _ext () < 0, performs step (4), otherwise, perform step (5);

(4) first, according to the rated capacity Q of each battery pack in t isolated island _{max, i}, state-of-charge Soc _ithe residual capacity Q in (t) and the first pond _{1, i}t (), utilizes following formula to try to achieve each accumulator lasting charge power P externally maximum acceptable in time interval Δ t _cmax(t):

Secondly, according to P _cmaxt the size of (), utilizes following formula to calculate the clean exchange power corresponding with each battery pack

$P_{ex}^i\left(t\right)=\frac{P_{cmax}^i\left(t\right)\&times;Q_{max,i}}{\underset{i=1}{\overset{N_b}{\&Sigma;}}{P}_{cmax}^i\left(t\right)\&times;Q_{max,i}}---\left(18\right)$

In formula, Q _{max, i}for the rated capacity of battery pack,

Finally, following formula is utilized to calculate the power of each battery pack actual absorption in Δ t

$P=\left\{\begin{array}{cc}\max (P_{c\max },P_{ex}^t){\mathrm{\&eta;}}_c& P_{ex}^t<0\\ \min (P_{d\max },P_{ex}^t)/{\mathrm{\&eta;}}_c& P_{ex}^t\&GreaterEqual;0\end{array}\right.---\left(19\right)$

And upgrade the state-of-charge Soc in t+ Δ t of each battery pack _ithe residual capacity Q in (t+ Δ t) and the first pond _{1, i}(t+ Δ t);

(5) first, according to the rated capacity Q of each battery pack in t isolated island _{max, i}, state-of-charge Soc _ithe residual capacity Q in (t) and the first pond _{1, i}(t), the peak power utilizing battery model to try to achieve each accumulator can externally to provide in time interval Δ t now, if size meet the constraint of load summate condition, then perform step (6); Otherwise, perform step (7);

(6) first, according to size, utilize following formula to calculate the clean exchange power corresponding with each battery pack

$P_{ex}^i\left(t\right)=\frac{P_{d\max }^i\left(t\right)\&times;Q_{\max ,i}}{\underset{i=1}{\overset{N_b}{\&Sigma;}}{P}_{d\max }^i\left(t\right)\&times;Q_{\max ,i}}---\left(20\right)$

Secondly, formula is utilized $P=\left\{\begin{array}{cc}\max (P_{c\max },P_{ex}^t){\mathrm{\&eta;}}_c& P_{ex}^t<0\\ \min (P_{d\max },P_{ex}^t)/{\mathrm{\&eta;}}_c& P_{ex}^t\&GreaterEqual;0\end{array}\right.$ Calculate the power of the actual release of each battery pack in Δ t and upgrade the state-of-charge Soc in t+ Δ t of each battery pack _ithe residual capacity Q in (t+ Δ t) and the first pond _{1, i}(t+ Δ t); Finally, time t=t is made _st+ Δ t, and return step (3);

(7) by all load point be retained, total electricity W _kreduction state X (k) of minimum load point is set as 0, as time t < t _endtime, return step (2); As time t=t _endtime, load summate flow process terminates, and records reduction state X (k) of final each load point.

In above-mentioned flow process, owing to comprising multiple battery pack in isolated island, the rated capacity of each battery pack and may be different at the initial state-of-charge of synchronization, in order to balance the charge status between each battery pack, in each moment of islet operation, each battery pack needs the power of release according to its current time the distribution that is in proportion, need absorb power then press the distribution that is in proportion.In addition, for the load point in isolation isolated island district, downstream and seamless isolated island district, downstream, if a certain load point is all cut down within two parts isolated island time, think that its frequency of power cut is 1, and power off time adds up.

Adopting the RBTSBus6 feeder system of improvement below, is original distribution network with main feeder F4, provides an example.

By Fig. 2 interior joint 13 place access distributed power source, design parameter is as follows:

1) blower fan: separate unit blower fan rated power 335kW; Incision wind speed 2.5m/s; Wind rating 12.5m/s; Cut-out wind speed 25m/s; Fitting parameter A, B, C are respectively-39.58,6.37,2.02; Mean wind speed is 19.56m/s, and wind speed profile standard deviation is 10.06m/s.

2) photovoltaic array: parameter R _cand G _stdbe respectively 0.15kW/m ²and 1kW/m ².

3) accumulator: every block battery rating 3000Ah, rated voltage 2V (6kWh); Parameter c=0.317, α=1, k=1.22, η _c=η _d=0.927, I _max=610A.

4) element failure rate: feeder fault rate is 0.05 times/year × km, distribution transforming failure rate is 0.015 times/year, and switch fault rate is 0.006 times/year, and mean repair time is 5h, obeys index distribution.Fault isolation and load turn the band time and get steady state value 1h.Separate unit fan trouble state probability P _d=7.3%; Photovoltaic array is identical with the state model parameter of battery pack, malfunction probability P _d=3.2%, derate state probability P _e=5%.

Adopt the construction method of the present invention heuristic load summate model in the isolated island of active power distribution network to analyze above-mentioned example, Fig. 3 represents and does not consider load summate strategy and the contrast situation of annual power off time considering load in load summate strategy two kinds of situations containing distributed power source containing distributed power source.

The validity of the construction method of heuristic load summate model in the isolated island of active power distribution network is described by the length analyzing contrast annual power off time.

Shown by analysis diagram, the total electricity of load of the load block that load point 19,20,21 forms is larger, finally just excise this part load when that is carrying out load summate, therefore after considering load summate strategy, the reliability of these 3 points is slightly higher than and does not consider load summate strategy.And the total electricity of load of the load block be made up of load point 22,23 is less, therefore when considering load summate strategy, this reliability index of 2 is than bigger than normal when not considering.When can find out distributed power source undercapacity in isolated island, carry out load summate by the present invention, can power supply and demand balance be ensured, make important large-scale load be able to normal operation.Under the background that current distributed power source capacity expands rapidly, fully, the islet operation characteristic of Appropriate application distributed power source, release power supply potential, for energy-saving and emission-reduction, rationally carries out distribution network planning and has certain directive significance.

Claims (7)

1. the construction method of heuristic load summate model in the isolated island of active power distribution network, is characterized in that, comprise the steps:

1) component models when building islet operation, comprises exert oneself model, photovoltaic of blower fan and to exert oneself model, load model and battery model;

2) load summate model is built;

3) solve isolated island internal loading and cut down model.

2. the construction method of heuristic load summate model in the isolated island of active power distribution network according to claim 1, is characterized in that, step 1) described in blower fan to exert oneself the structure of model, comprising:

(1) simulation of seasonal effect in time series arma modeling is adopted to produce the time series data of wind speed:

V _t＝μ _t+σ _ty _t(1)

y _t＝φ ₁y _t-1+φ ₂y _t-2+…φ _ly _t-l+…φ _ny _t-n+α _t-α _t-1θ ₁-α _t-2θ ₂-…α _t-sθ _s…-α _t-mθ _m(2)

In formula, V _tfor real-time wind speed; μ _tfor the mean value of historical wind speed data in assessment area, σ _tfor the standard deviation of historical wind speed distribution, y _tfor time series, φ _lfor autoregressive coefficient, l=1 ... n; θ _sfor running mean coefficient, s=1 ... m; α _tfor white noise coefficient, obedience average is 0, variance is independent normal distribution;

(2) set up blower fan to exert oneself model

$P_w=\left\{\begin{array}{cc}0,& 0\&le;V_t<V_{ci}\\ (A+B\&times;V_t+C\&times;V_t^2)P_r,& V_{ci}\&le;V_t<V_r\\ P_r,& V_r\&le;V_t\&le;V_{co}\\ 0,& V_t>V_{co}\end{array}\right.---\left(3\right)$

In formula, P _wfor exerting oneself in real time of blower fan, A, B, C are the coefficient of the fitting function of power curve non-linear partial, V _tbe the real-time air speed data of t hour, V _ci, V _rand V _cobe respectively the incision wind speed of blower fan, wind rating and cut-out wind speed, P _rfor the output rating of blower fan.

3. the construction method of heuristic load summate model in the isolated island of active power distribution network according to claim 1, is characterized in that, step 1) described in photovoltaic model of exerting oneself be:

$P_b=\left\{\begin{array}{cc}{P}_{sn}\left(G_{bt}^2/\left(G_{std}{R}_c\right)\right),& 0\&le;G_{bt}<R_c\\ P_{sn}\left(G_{bt}/G_{std}\right),& R_c\&le;G_{bt}<G_{std}\\ P_{sn},& G_{bt}\&GreaterEqual;G_{std}\end{array}\right.---\left(4\right)$

In formula, P _bfor exerting oneself in real time of photovoltaic, unit kW; P _snfor the rated power of photovoltaic, represent the power that unit light intensity can produce under standard test condition; G _stdfor specified intensity of illumination, unit is kW/m ²; R _cfor the light intensity of a certain certain strength, under this light intensity photovoltaic exert oneself to start with the relation of light intensity from non-linear become linear; G _btbe the real-time light intensity of t hour, unit kW/m ², G _btreal time sequence obtained by the sampling of the probability distribution statistical to history light intensity.

4. the construction method of heuristic load summate model in the isolated island of active power distribution network according to claim 1, it is characterized in that, step 1) described in load model be logical overladen Typical Year-all curves, week-curve and day-hour curve forms real-time load data and obtains:

L _t＝L _p×P _w×P _d×P _h(t)(5)

In formula, L _pfor year load peak, P _wfor corresponding with t hour year-all load curves in value, P _dfor corresponding with t hour week-daily load curve in value, P _h(t) be corresponding with t hour day-hour load curve in value.

5. the construction method of heuristic load summate model in the isolated island of active power distribution network according to claim 1, it is characterized in that, step 1) described in battery model be adopt two pool models of lead-acid accumulator, namely accumulator is divided into two ponds of mutually connecting, be wherein utilisable energy in the first pond, represent and can be converted into the energy of electric energy immediately; Be bound energy in second pond, represent and to be subject in pond chemical reaction velocity constraint and electric energy cannot be converted at once or be the energy of chemical energy by electric energy conversion;

Battery discharging model representation is:

P _dmax＝min(P _dkbm,P _misoc)η _d(6)

In formula, P _dmaxrepresent the maximum continuous discharge power that accumulator is external, P _dkbmrepresent the battery discharging restrain condition described by two pool models of accumulator, P _misocrepresent minimum capacity constraint, i.e. the minimum state-of-charge of accumulator, η _dfor discharging efficiency, P _dkbmand P _misoccomputing formula as follows:

$P_{dkbm}=\frac{{\mathrm{kQ}}_1{e}^{-k\mathrm{\&Delta;}t}+Qkc(1-e^{-k\mathrm{\&Delta;}t})}{1-e^{-k\mathrm{\&Delta;}t}+c(k\mathrm{\&Delta;}t-1+e^{-k\mathrm{\&Delta;}t})}---\left(7\right)$

P _misoc＝Q _max(Soc _st-Soc _min)/△t(8)

Wherein Q is the total surplus capacity of accumulator before electric discharge, Q ₁for the residual capacity in front first pond of discharging, Soc _minfor the minimum state-of-charge constraint of accumulator, Soc _stfor the state-of-charge of the front accumulator that charges, △ t is any time period, Q _maxbe the total volume in two ponds, i.e. the rated capacity (kWh) of accumulator; C is the ratio of the first capacity and total volume, represents the capacity ratio in two ponds; K is constant (1/h), in order to the conversion rate of Characterization Energy between Liang Chi.

Charge in batteries model representation is:

P _cmax＝max(P _ckbm,-P _mcr,-P _mcc,-P _masoc)/η _c(9)

In formula, P _cmaxfor the maximum acceptable lasting charge power that accumulator is external, P _ckbmrepresent the charge in batteries restrain condition described by two pool models, P _mcrrepresent the maximum charge rate constraint of accumulator, P _mccrepresent the maximum permission charging current constraint of accumulator, P _masocrepresent max cap. constraint, i.e. the most highly charged state of accumulator, η _cfor charge efficiency.The computing method of each constraint are as follows:

$P_{ckbm}=\frac{-{\mathrm{kcQ}}_{max}+{\mathrm{kQ}}_1{e}^{-k\mathrm{\&Delta;}t}+Qkc(1-e^{-k\mathrm{\&Delta;}t})}{1-e^{-k\mathrm{\&Delta;}t}+c(k\mathrm{\&Delta;}t-1+e^{-k\mathrm{\&Delta;}t})}---\left(10\right)$

P _mcr＝(1-e ^-α△t)(Q _max-Q)/△t(11)

P _mcc＝I _maxV _nom/1000(12)

P _masoc＝Q _max(Soc _max-Soc _st)/△t(13)

Wherein, I _maxfor the maximum permission charging current of accumulator, V _nomfor accumulator rated operational voltage, Soc _maxfor the most highly charged state constraint.

6. the construction method of heuristic load summate model in the isolated island of active power distribution network according to claim 1, it is characterized in that, step 2) described in structure load summate model, first the condition of load summate is carried out in setting: the total load be more than or equal in this moment island of exerting oneself of the peak power that during islet operation, in island, all battery pack can discharge and all distributed power sources, and mathematic(al) representation is as follows:

$\begin{array}{c}\underset{i=1}{\overset{N_b}{\&Sigma;}}{P}_{d\max }^i\left(t\right)+\underset{j=1}{\overset{N_{DG}}{\&Sigma;}}{P}_{DG}^j\left(t\right)\&GreaterEqual;S_l\left(t\right)+\underset{k=1}{\overset{N_L}{\&Sigma;}}{L}_t^k\left(t\right)X\left(k\right),t\&Element;\&lsqb;t_{st},t_{end}\&rsqb;\\ when{S}_l\left(t\right)+\underset{k=1}{\overset{N_L}{\&Sigma;}}{L}_t^k\left(t\right)X\left(k\right)-\underset{j=1}{\overset{N_{DG}}{\&Sigma;}}{P}_{DG}^j\left(t\right)\&GreaterEqual;0\end{array}---\left(14\right)$

In formula, for jth distributed power source exerting oneself in real time in t, blower fan by blower fan exert oneself model calculate, photovoltaic by photovoltaic exert oneself model calculate, N _dGfor the sum of distributed power source; for the Real-time Load of a kth load point, calculated by load model, N _lfor the sum of load point; for the peak power that t i-th battery pack can externally provide in time interval △ t, calculated by battery model, N _bfor total number of battery pack in isolated island, S _lt () is the active loss in isolated island, t _stfor isolated island forms the moment, t _endfor islet operation finish time, time interval △ t gets 1 hour, and X (k) is the reduction state of a kth load point.

If do not meet the constraint of the condition of formula load summate, just need the load cutting down a part of load point, to guarantee the power balance in isolated island, the objective function of load summate is:

$\begin{array}{cc}\max & \underset{t=t_{st}}{\overset{t_{end}}{\&Sigma;}}\underset{k=1}{\overset{N_L}{\&Sigma;}}{L}_t^k\left(t\right)X\left(k\right)\end{array}---\left(15\right)$

In formula, X (k) is the reduction state of the kth load point in the condition of load summate, and X (k)=0 represents load point k and cut down, and X (k)=1 represents load point k and is retained;

The condition of load summate and the objective function of load summate constitute the load summate model in isolated island, and the objective function of load summate is the objective function of model, and the condition of load summate is constraint condition.

7. the construction method of heuristic load summate model in the isolated island of active power distribution network according to claim 1, is characterized in that, step 3) described in solve isolated island internal loading cut down model, comprise the steps:

(1) reduction state X (k) setting each load point is 1, and calculates the total electricity W of each load point during isolated island according to the following formula _k,

$W_k=\underset{t=t_{st}}{\overset{t_{end}}{\&Sigma;}}{L}_t^k\left(t\right)---\left(16\right)$

(2) time t=t is made _st;

(3) the clean exchange power P in t isolated island is calculated according to the following formula _ex(t),

$P_{ex}\left(t\right)=\underset{k=1}{\overset{N_L}{\&Sigma;}}{L}_t^k\left(t\right)X\left(k\right)+S_l\left(t\right)-\underset{j=1}{\overset{N_{DG}}{\&Sigma;}}{P}_{DG}^j\left(t\right)---\left(17\right)$

If P _ext () <0, performs step (4), otherwise, perform step (5);

(4) first, according to the rated capacity Q of each battery pack in t isolated island _{max, i}, state-of-charge Soc _ithe residual capacity Q in (t) and the first pond _{1, i}t (), utilizes following formula to try to achieve each accumulator lasting charge power P externally maximum acceptable in time interval △ t _cmax(t);

Secondly, according to P _cmaxt the size of (), utilizes following formula to calculate the clean exchange power corresponding with each battery pack

$P_{ex}^i\left(t\right)=\frac{P_{cmax}^i\left(t\right)\&times;Q_{max,i}}{\underset{i=1}{\overset{N_b}{\&Sigma;}}{P}_{cmax}^i\left(t\right)\&times;Q_{\max ,i}}---\left(18\right)$

In formula, Q _{max, i}for the rated capacity of battery pack,

Finally, following formula is utilized to calculate the power of each battery pack actual absorption in △ t

$P=\left\{\begin{array}{cc}max\left(P_{c\max },P_{ex}^t\right){\mathrm{\&eta;}}_c& P_{ex}^t<0\\ min\left(P_{d\max },P_{ex}^t\right)/{\mathrm{\&eta;}}_d& P_{ex}^t\&GreaterEqual;0\end{array}\right.---\left(19\right)$

And upgrade the state-of-charge Soc in t+ △ t of each battery pack _ithe residual capacity Q in (t+ △ t) and the first pond _{1, i}(t+ △ t);

(5) first, according to the rated capacity Q of each battery pack in t isolated island _{max, i}, state-of-charge Soc _ithe residual capacity Q in (t) and the first pond _{1, i}(t), the peak power utilizing battery model to try to achieve each accumulator can externally to provide in time interval △ t now, if size meet the constraint of load summate condition, then perform step (6); Otherwise, perform step (7);

(6) first, according to size, utilize following formula to calculate the clean exchange power corresponding with each battery pack

$P_{\mathrm{ex}}^i\left(t\right)=\frac{P_{dmax}^i\left(t\right)\&times;Q_{max,i}}{\underset{i=1}{\overset{N_b}{\&Sigma;}}{P}_{d\max }^i\left(t\right)\&times;Q_{\max ,i}}---\left(20\right)$

Secondly, formula is utilized $P=\left\{\begin{array}{cc}max\left(P_{c\max },P_{ex}^t\right){\mathrm{\&eta;}}_c& P_{ex}^t<0\\ min\left(P_{d\max },P_{ex}^t\right)/{\mathrm{\&eta;}}_d& P_{ex}^t\&GreaterEqual;0\end{array}\right.$ Calculate the power of the actual release of each battery pack in △ t and upgrade the state-of-charge Soc in t+ △ t of each battery pack _ithe residual capacity Q in (t+ △ t) and the first pond _{1, i}(t+ △ t); Finally, time t=t is made _st+ △ t, and return step (3);

(7) by all load point be retained, total electricity W _kreduction state X (k) of minimum load point is set as 0, as time t<t _endtime, return step (2); As time t=t _endtime, load summate flow process terminates, and records reduction state X (k) of final each load point.