CN103326387B - Source network coordinated dispatching method reducing wind electricity dispatching risks by means of stored energy - Google Patents

Abstract

The invention relates to a source network coordinated dispatching method reducing wind electricity dispatching risks by means of stored energy. The source network coordinated dispatching method is characterized by comprising the steps of prediction of a wind electricity power probability interval, determination of a wind electricity dispatching scheme, allocation of storage energy system capacity, and the like, wherein the energy storage system has time shifting capacity on power and energy, predication errors of the wind electricity power can be effectively responded, the wind electricity dispatching risks are reduced, safe operation of a power grid is guaranteed, available space of the power grid can be utilized to the maximum degree on the premise that safety of the power grid is guaranteed, the power generation amount of a wind turbine generator is improved, wind electricity abandoned wind is reduced, and economic benefits and environmental benefits of wind energy utilization are improved.

Description

A kind of source net coordinated scheduling method utilizing energy storage to reduce wind-powered electricity generation schedule risk

Technical field

The invention belongs to technical field of wind power generation, is a kind of source net coordinated scheduling method utilizing energy storage to reduce wind-powered electricity generation schedule risk.

Background technology

In recent years, global wind power generation development is very fast.China in recent years installed capacity of wind-driven power with close to 100% speed increment, built up a collection of large-scale wind power base, these wind power base are mainly carried in the mode of networking operation and are distributed electric energy.

At the initial stage of Wind Power Development, it is less that wind-powered electricity generation accounts for the electrical network proportion that always generates electricity, and the access of wind-powered electricity generation is very little on operation of power networks impact.After large-scale wind power base access electrical network, the proportion that installed capacity of wind-driven power accounts for system total installed capacity is larger.Due to wind power can not Accurate Prediction, for not affecting electric power netting safe running, dispatching of power netwoks person is usually by the conservative manner determination wind-powered electricity generation scheduling networking capacity considering largest prediction error, and this Wind turbines that causes networking abandons wind in a large number or newly-built Wind turbines can not be incorporated into the power networks.The main wind power base such as Gansu, the Inner Mongol and Jilin is rationed the power supply and is abandoned wind ratio all more than 20% according to statistics, within 2012, the whole nation affects wind power generation capacity reach 20,000,000,000 kWh because limiting grid-connected, account for 20% of whole wind power generation capacity (1,004 hundred million kWh), the electricity of loss amounts to standard coal more than 6,700,000 tons, and wind-powered electricity generation enterprise is because factor loss of rationing the power supply is more than 10,000,000,000 yuans.

Therefore, be necessary to study and improve wind-powered electricity generation scheduling networking scale, reduce the source net coordinated scheduling method that wind-powered electricity generation abandons wind, with receiving wind-powered electricity generation as much as possible under the prerequisite realizing ensureing power grid security.

Summary of the invention

Technical problem to be solved by this invention is: the large-scale wind power existed for China's large-scale wind power networking operation abandons wind problem, propose a kind of source net coordinated scheduling method utilizing energy storage to reduce wind-powered electricity generation schedule risk, thus on the basis realizing guarantee electric power netting safe running, improve wind-powered electricity generation receiving scale.

Solving the technical scheme that its technical problem adopts is: a kind of source net coordinated scheduling method utilizing energy storage to reduce wind-powered electricity generation schedule risk, and it is characterized in that, it comprises the following steps:

1) the probability interval prediction of wind power

Wind power prediction is the basis solving wind-powered electricity generation scheduling problem, in order to ensure the safe operation of electrical network, the maximum of arranged wind-powered electricity generation scheduling networking capacity must be made can not break through the utilized space of electrical network by generated output; For this reason, need the wind-powered electricity generation maximum output value predicting each scheduling each scheduling slot of day, and guarantee that it does not break through the safe operation limit of electrical network; The wind power maximum of future scheduling period employing formula (1) is predicted:

$P_{k,\mathrm{Day}\_N}^{*\mathrm{forcast}}={\mathrm{\φ}}_0+{\mathrm{\φ}}_1\max P_k^{*\mathrm{his}}+{\mathrm{\φ}}_2{\max P}_{k-1}^{*\mathrm{his}}+\·\·\·+{\mathrm{\φ}}_p{\max P}_{k-p}^{*\mathrm{his}}+{\mathrm{\ϵ}}_t---\left(1\right)$

Wherein, represent the wind power prediction value of scheduling slot, for the wind power maximum output value of history scheduling slot, p represents the exponent number of forecast model, φ ₀for constant term; φ ₁, φ ₂, φ _pfor autoregressive coefficient, ε _tfor random disturbances amount;

The predicated error of wind power is the difference between actual wind power value of a certain moment and prediction, namely has:

$e_{i,k}=P_{i,k}^{\mathrm{real}}-P_{i,k}^{\mathrm{forecast}}---\left(2\right)$

Wherein e _i,kfor the predicated error of history i-th day kth scheduling slot, P _i,k ^realfor the actual maximum wind performance number of history i-th day kth scheduling slot, for history i-th day kth scheduling slot prediction wind-powered electricity generation maximum output value;

If Ω (j) represent t before the set of nearest n prediction error value, if the probability of each element is 1/n in set omega (j), then the empirical distribution function of Ω (j) be described below:

${\hat{F}}_t\left(\mathrm{\ξ}\right)=\frac{1}{n}\mathrm{Num}\{e_i\≤\mathrm{\ξ}|e_i\∈\mathrm{\Ω}\left(j\right),i=\mathrm{1,2}\·\·\·,n\}---\left(3\right)$

Wherein: function Num is used for asking in set omega (j) number of elements meeting specified criteria, ξ is a stochastic variable of predicated error, represent that predicated error is less than the probability of ξ, n is the sum of history data set, Ω (j) represent t before the set of nearest n prediction error value, e _irepresent the predicated error of i-th scheduling slot;

Suppose for empirical distribution function inverse function, then a α probabilistic forecasting interval of actual value is:

$[P_{t|t+k}^{\mathrm{real}}+{\hat{G}}_t\left({\mathrm{\α}}_1\right),P_{t|t+k}^{\mathrm{real}}+{\hat{G}}_t\left({\mathrm{\α}}_2\right)]---\left(4\right)$

Wherein for the wind power actual value of a kth scheduling slot, α ₁lower limit for wind power probability interval:

α ₁=(1-α)/2, α ₂higher limit for wind power probability interval: α ₂=1-(1-α)/2, for empirical distribution function inverse function;

2) determination of wind-powered electricity generation scheduling scheme

Receive wind-powered electricity generation to realize maximum-norm, its principle makes to dispatch the wind-powered electricity generation networked can utilize electrical network that space can be utilized not jeopardize again the safety of electrical network to greatest extent by generated output;

Consider that wind power prediction can exist larger predicated error, the actual value can not getting rid of wind power is greater than the situation appearance of predicted value, and this installed capacity that preliminary dispatching office will be caused to obtain exceedes the receiving limit of electrical network, jeopardizes the safe operation of electrical network;

For this reason, α probability interval is adopted to carry out the arrangement of wind-powered electricity generation scheduling scheme, to realize reducing the uncertain risk brought to operation of power networks of wind power prediction to greatest extent, so the k period dispatches the installed capacity of wind-driven power networked and is:

$\begin{array}{c}{C}_k^{\mathrm{sp},\mathrm{pre}}=\frac{C_{\mathrm{Basc}}^{*}}{{\stackrel{\‾}{P}}_{k,\mathrm{Day}\_Q+1}^{*\mathrm{forcast}}\left(\mathrm{\α}\right)}\×P_k^{\mathrm{space}}\\ \mathrm{st}.P_k^{\mathrm{space}}\≤C_k^{\mathrm{sp},\mathrm{pre}}\≤C_{\mathrm{\Σ},w}\end{array}---\left(5\right)$

Wherein: C _k ^spacerepresent the scheduling networking capacity of wind-powered electricity generation; represent the higher limit of wind power α probability interval prediction, calculated by formula (4), C ^* _baserepresent the reference capacity for carrying out wind power prediction, P _k ^spacerepresent the utilized space of electrical network, in formula (5), constraints limits scheduling networking installed capacity of wind-driven power C _k ^{sp, pre}the electrical network being greater than scheduling slot can utilize space P _k ^space, be less than intrasystem total installed capacity of wind-driven power C _{Σ, w};

3) energy storage system capacity configuration

Formula (5) is the source net coordinated scheduling scheme based on the prediction of wind-powered electricity generation probability interval, but for each scheduling day, adopt α probability interval forecast interval can not ensure that wind power prediction is worth 100% probability and drops in this interval, this can cause the wind-powered electricity generation of arranged scheduling scheme can generated output cross the safe operation limit of electrical network situation occur, jeopardize the safe operation of electrical network; Therefore, need to utilize energy-storage system to carry out charging to ensure the safe operation of electrical network; Corresponding energy-storage system charge and discharge control strategy is as follows:

$\left\{\begin{array}{c}{E}_k^{\mathrm{ch}\arg e}=E_0+\underset{i=1}{\overset{N}{\mathrm{\Σ}}}(P_{\mathrm{ref}}^i\×\mathrm{\ΔT}\×{\mathrm{\η}}_{\mathrm{ch}\arg e})\\ E_k^{\mathrm{disch}\arg e}=E_0+\underset{i=1}{\overset{N}{\mathrm{\Σ}}}(P_{\mathrm{ref}}^i\×\mathrm{\ΔT}/{\mathrm{\η}}_{\mathrm{disch}\arg e})\end{array}\right.---\left(6\right)$

Wherein E _k ^charge, E _k ^discharge, E ₀be expressed as energy-storage system rechargeable energy, discharge energy and initial state-of-charge, N is the charge and discharge control number of times of each scheduling slot energy-storage system, and Δ T is the basic controlling cycle of energy-storage system, and it is relevant with sampling time interval; η _charge, η _dischargerepresent the charge and discharge efficiency of energy-storage system respectively, P ⁱ _reffor i moment energy-storage system charge power reference value, its computing formula is as follows:

$P_{\mathrm{ref}}^i=P_{\mathrm{real}}^i-P_k^{\mathrm{space}}---\left(7\right)$

Wherein P ⁱ _refrepresent the i-th moment energy-storage system charge power reference value, P ⁱ _realrepresent the actual value of the i-th moment wind power, P _k ^spacerepresent that the electrical network of scheduling slot can utilize space;

In order to ensure that the wind-powered electricity generation scheduling networking scale arranged in any scheduling slot is no more than the safe operation limit of electrical network, then energy storage system capacity need be equipped with by the maximum of scheduling slot each in dispatching cycle, and its computing formula is as follows:

$C_w=\max \left(\underset{n=1}{\overset{N}{\∪}}{\left\{E\right\}}_{k,\mathrm{day}\_N}\right)---\left(8\right)$

Wherein N is total scheduling day, C _wfor the energy storage system capacity of required outfit, { E} _{k, day_N}represent the rechargeable energy of energy-storage system at a N days kth scheduling slot.

A kind of source net coordinated scheduling method utilizing energy storage to reduce wind-powered electricity generation schedule risk proposed by the invention, the large-scale wind power existed mainly for China's large-scale wind power networking operation abandons wind problem, wind-powered electricity generation scheduling decision is carried out in the prediction of its probability interval based on wind power, and utilize energy-storage system tackle wind-powered electricity generation exist schedule risk, can realize ensure electric power netting safe running basis on improve wind-powered electricity generation receive scale.

Accompanying drawing explanation

Schematic diagram is sent in the binding of Fig. 1 wind fire outside;

The electrical network that Fig. 2 dispatches each scheduling slot of day can utilize space;

The predicated error of Fig. 3 wind power;

80% probability interval of Fig. 4 wind power predicts the outcome;

The wind-powered electricity generation scheduling networking capacity of each scheduling slot of Fig. 5;

The wind power generation capacity of each scheduling slot of Fig. 6;

The rechargeable energy of Fig. 7 day part energy-storage system;

The electrical network space peak use rate of each scheduling slot of Fig. 8.

In figure: shade 1 represents that 80% probabilistic forecasting of wind power is interval, curve 2 represents the actual wind-powered electricity generation maximum output value of each scheduling slot, curve 3 represents the wind-powered electricity generation maximum output predicted value of each scheduling slot, curve 4 represents the wind-powered electricity generation scheduling networking capacity of each scheduling slot of the inventive method, curve 5 represents the wind-powered electricity generation scheduling networking capacity of each scheduling slot of conventional method, curve 6 represents the scheduling day wind power generation capacity of the inventive method, curve 7 represents the scheduling day wind power generation capacity of conventional method, curve 8 represents the electrical network space peak use rate of the inventive method, curve 9 represents the electrical network space peak use rate of conventional method.

Embodiment

1) the probability interval prediction of wind power

Wind power prediction is the basis solving wind-powered electricity generation scheduling problem, in order to ensure the safe operation of electrical network, the maximum of arranged wind-powered electricity generation scheduling networking capacity must be made can not break through the utilized space of electrical network by generated output; For this reason, need the wind-powered electricity generation maximum output value predicting each scheduling each scheduling slot of day, and guarantee that it does not break through the safe operation limit of electrical network; The wind power maximum of future scheduling period employing formula (1) is predicted:

$P_{k,\mathrm{Day}\_N}^{*\mathrm{forcast}}={\mathrm{\φ}}_0+{\mathrm{\φ}}_1\max P_k^{*\mathrm{his}}+{\mathrm{\φ}}_2{\max P}_{k-1}^{*\mathrm{his}}+\·\·\·+{\mathrm{\φ}}_p{\max P}_{k-p}^{*\mathrm{his}}+{\mathrm{\ϵ}}_t---\left(1\right)$

Wherein, represent the wind power prediction value of scheduling slot, for the wind power maximum output value of history scheduling slot, p represents the exponent number of forecast model, φ ₀for constant term; φ ₁, φ ₂, φ _pfor autoregressive coefficient, ε _tfor random disturbances amount;

The predicated error of wind power is the difference between actual wind power value of a certain moment and prediction, namely has:

$e_{i,k}=P_{i,k}^{\mathrm{real}}-P_{i,k}^{\mathrm{forecast}}---\left(2\right)$

Wherein e _i,kfor the predicated error of history i-th day kth scheduling slot, P _i,k ^realfor the actual maximum wind performance number of history i-th day kth scheduling slot, for history i-th day kth scheduling slot prediction wind-powered electricity generation maximum output value;

If Ω (j) represent t before the set of nearest n prediction error value, if the probability of each element is 1/n in set omega (j), then the empirical distribution function of Ω (j) be described below:

${\hat{F}}_t\left(\mathrm{\ξ}\right)=\frac{1}{n}\mathrm{Num}\{e_i\≤\mathrm{\ξ}|e_i\∈\mathrm{\Ω}\left(j\right),i=\mathrm{1,2}\·\·\·,n\}---\left(3\right)$

Wherein: function Num is used for asking in set omega (j) number of elements meeting specified criteria, ξ is a stochastic variable of predicated error, represent that predicated error is less than the probability of ξ, n is the sum of history data set, Ω (j) represent t before the set of nearest n prediction error value, e _irepresent the predicated error of i-th scheduling slot;

Suppose for empirical distribution function inverse function, then a α probabilistic forecasting interval of actual value is:

$[P_{t|t+k}^{\mathrm{real}}+{\hat{G}}_t\left({\mathrm{\α}}_1\right),P_{t|t+k}^{\mathrm{real}}+{\hat{G}}_t\left({\mathrm{\α}}_2\right)]---\left(4\right)$

Wherein for the wind power actual value of a kth scheduling slot, α ₁lower limit for wind power probability interval:

α ₁=(1-α)/2, α ₂higher limit for wind power probability interval: α ₂=1-(1-α)/2, for empirical distribution function inverse function;

2) determination of wind-powered electricity generation scheduling scheme

Receive wind-powered electricity generation to realize maximum-norm, its principle makes to dispatch the wind-powered electricity generation networked can utilize electrical network that space can be utilized not jeopardize again the safety of electrical network to greatest extent by generated output;

Consider that wind power prediction can exist larger predicated error, the actual value can not getting rid of wind power is greater than the situation appearance of predicted value, and this installed capacity that preliminary dispatching office will be caused to obtain exceedes the receiving limit of electrical network, jeopardizes the safe operation of electrical network;

For this reason, α probability interval is adopted to carry out the arrangement of wind-powered electricity generation scheduling scheme, to realize reducing the uncertain risk brought to operation of power networks of wind power prediction to greatest extent, so the k period dispatches the installed capacity of wind-driven power networked and is:

$\begin{array}{c}{C}_k^{\mathrm{sp},\mathrm{pre}}=\frac{C_{\mathrm{Basc}}^{*}}{{\stackrel{\‾}{P}}_{k,\mathrm{Day}\_Q+1}^{*\mathrm{forcast}}\left(\mathrm{\α}\right)}\×P_k^{\mathrm{space}}\\ \mathrm{st}.P_k^{\mathrm{space}}\≤C_k^{\mathrm{sp},\mathrm{pre}}\≤C_{\mathrm{\Σ},w}\end{array}---\left(5\right)$

Wherein: C _k ^spacerepresent the scheduling networking capacity of wind-powered electricity generation; represent the higher limit of wind power α probability interval prediction, calculated by formula (4), C ^* _baserepresent the reference capacity for carrying out wind power prediction, P _k ^spacerepresent the utilized space of electrical network, in formula (5), constraints limits scheduling networking installed capacity of wind-driven power C _k ^{sp, pre}the electrical network being greater than scheduling slot can utilize space P _k ^space, be less than intrasystem total installed capacity of wind-driven power C _{Σ, w};

3) energy storage system capacity configuration

Formula (5) is the source net coordinated scheduling scheme based on the prediction of wind-powered electricity generation probability interval, but for each scheduling day, adopt α probability interval forecast interval can not ensure that wind power prediction is worth 100% probability and drops in this interval, this can cause the wind-powered electricity generation of arranged scheduling scheme can generated output cross the safe operation limit of electrical network situation occur, jeopardize the safe operation of electrical network; Therefore, need to utilize energy-storage system to carry out charging to ensure the safe operation of electrical network; Corresponding energy-storage system charge and discharge control strategy is as follows:

$\left\{\begin{array}{c}{E}_k^{\mathrm{ch}\arg e}=E_0+\underset{i=1}{\overset{N}{\mathrm{\Σ}}}(P_{\mathrm{ref}}^i\×\mathrm{\ΔT}\×{\mathrm{\η}}_{\mathrm{ch}\arg e})\\ E_k^{\mathrm{disch}\arg e}=E_0+\underset{i=1}{\overset{N}{\mathrm{\Σ}}}(P_{\mathrm{ref}}^i\×\mathrm{\ΔT}/{\mathrm{\η}}_{\mathrm{disch}\arg e})\end{array}\right.---\left(6\right)$

Wherein E _k ^charge, E _k ^discharge, E ₀be expressed as energy-storage system rechargeable energy, discharge energy and initial state-of-charge, N is the charge and discharge control number of times of each scheduling slot energy-storage system, and Δ T is the basic controlling cycle of energy-storage system, and it is relevant with sampling time interval; η _charge, η _dischargerepresent the charge and discharge efficiency of energy-storage system respectively, P ⁱ _reffor i moment energy-storage system charge power reference value, its computing formula is as follows:

$P_{\mathrm{ref}}^i=P_{\mathrm{real}}^i-P_k^{\mathrm{space}}---\left(7\right)$

Wherein P ⁱ _refrepresent the i-th moment energy-storage system charge power reference value, P ⁱ _realrepresent the actual value of the i-th moment wind power, P _k ^spacerepresent that the electrical network of scheduling slot can utilize space;

In order to ensure that the wind-powered electricity generation scheduling networking scale arranged in any scheduling slot is no more than the safe operation limit of electrical network, then energy storage system capacity need be equipped with by the maximum of scheduling slot each in dispatching cycle, and its computing formula is as follows:

$C_w=\max \left(\underset{n=1}{\overset{N}{\∪}}{\left\{E\right\}}_{k,\mathrm{day}\_N}\right)---\left(8\right)$

Wherein N is total scheduling day, C _wfor the energy storage system capacity of required outfit, { E} _{k, day_N}represent the rechargeable energy of energy-storage system at a N days kth scheduling slot.

Utilize accompanying drawing below and implement example and a kind of source net coordinated scheduling method utilizing energy storage to reduce wind-powered electricity generation schedule risk of the present invention specifically to be implemented and validity is evaluated;

1. the evaluation index of source net harmony

The wind-powered electricity generation scheduling scheme that source net is coordinated calculates based on historical data.In fact, when formulating scheduling scheme, the wind-powered electricity generation change of scheduling day is unknown.So, the much degree of the scheduling scheme formulated improve source net harmony, need to evaluate.This evaluation must adopt the wind-powered electricity generation data (instead of history numeric field data) of scheduling day reality.As example, the inspection of rolling can be carried out on history data set.

Investigation source net harmony mainly considers two aspects.

From the angle of " net ", mainly investigate the utilance in electrical network space.Because wind-powered electricity generation is change, a macroscopical investigation can only be done.The present invention only acts on certain period maximum wind power that scheduling networking installed capacity sends account for the percentage in this period electrical network space as evaluation index to dispatch day wind speed.Namely have, k period electrical network space peak use rate:

${\mathrm{\η}}_{\mathrm{grid}.\max }=\frac{\max \{\max {\left\{P^{*\mathrm{disp}}\right\}}_k\×C_k^{\mathrm{sp}}\}}{P_k^{\mathrm{space}}\×1000}\×100\%---\left(9\right)$

η _grid.max∈ (0,1), this value is larger, shows that electrical network space obtains utilizing more fully.

From the angle of " source " source net harmony, paper examines is networking installed capacity and wind power generation capacity on average.

2. example condition

This example is taken from certain wind fire binding and is sent scene outside, and on transmission cross-section, transmission power is restricted.Accompanying drawing 1 delivers to major network schematic diagram outside concentrating for certain province 4000MW installed capacity wind farm group and installation 2000MW normal power supplies bundle.The maximum transmission capacity P of the power Transmission Corridor PTC (Public Transmission Corridors) of normal power supplies (Thermal Power Plant) and wind farm group in figure _limit ^trans=3500MW.

Accompanying drawing 2 is that the electrical network of scheduling day day part can utilize space result of calculation.

Utilize formula (1) to carry out the prediction of wind power maximum to formula (4), the wind power prediction error obtained as shown in Figure 3; 80% probabilistic forecasting compartmental results as shown in Figure 4;

3. the validation verification of example calculating and method

In order to verify the validity of scheduling scheme proposed by the invention, example calculate 20 scheduling day each scheduling slot wind-powered electricity generation scheduling networking capacity, energy output and electrical network space peak use rate.Consider the traditional scheduler scheme of largest prediction error as what compare.

According to the electrical grid transmission space of 24 periods and the wind power 80% forecast interval maximum of day part, calculate 20 days day parts with networking installed capacity as shown in Figure 5 by formula (5).

Source net coordinated scheduling scheme with contrast the wind power generation capacity of scheme day part as shown in Figure 6.

From accompanying drawing 5,6, source net coordinated scheduling method significantly improves networking installed capacity and the wind power generation capacity of day part.In some circumstances, maximum scheduling networking installed capacity reaches 4000MW, this means all accessible electrical network of whole installed capacity.

The energy-storage system charging capacity of 20 scheduling each scheduling slots of day is calculated as schemed shown in attached 7 according to formula (6), formula (7).The energy storage system capacity that can calculate the required outfit when wind-powered electricity generation scheduling decision is carried out in employing 80% probabilistic forecasting interval according to formula (8) is 118MW.h.

Calculate according to formula (9) the electrical network space peak use rate that 20 are dispatched in a few days each scheduling slot, accordingly result as shown in Figure 8.From accompanying drawing 8, source net coordinated scheduling scheme significantly improves the peak use rate in day part electrical network space, and in 480 scheduling slots, do not occur that arranged wind-powered electricity generation exceeds the situation in electrical network space.

Table 1 compared for a kind of overall performance utilizing the source net coordinated scheduling method of energy storage reduction wind-powered electricity generation schedule risk to contrast with traditional conventional method of the present invention for all dispatching networking capacity, gross generation and electrical network space peak use rate mean value three indexs from 20 scheduling day breeze level.

Table 1

As seen from table: a kind of source net coordinated scheduling method utilizing energy storage to reduce wind-powered electricity generation schedule risk proposed make wind-powered electricity generation on average dispatch networking capacity, wind-powered electricity generation gross generation and electrical network space availability ratio comparatively conventional method improve 28.67%, 79.2%, 76.6% respectively.

Claims (1)

1. utilize energy storage to reduce a source net coordinated scheduling method for wind-powered electricity generation schedule risk, it is characterized in that, it comprises the following steps:

1) the probability interval prediction of wind power

Wind power prediction is the basis solving wind-powered electricity generation scheduling problem, in order to ensure the safe operation of electrical network, the maximum of arranged wind-powered electricity generation scheduling networking capacity must be made can not break through the utilized space of electrical network by generated output; For this reason, need the wind-powered electricity generation maximum output value predicting each scheduling each scheduling slot of day, and guarantee that it does not break through the safe operation limit of electrical network; The wind power maximum of future scheduling period employing formula (1) is predicted:

$P_{k,\mathrm{Day}\_N}^{*\mathrm{forcast}}={\mathrm{\φ}}_0+{\mathrm{\φ}}_1\max P_k^{*\mathrm{his}}+{\mathrm{\φ}}_2\max P_{k-1}^{*\mathrm{his}}+...+{\mathrm{\φ}}_p\max P_{k-p}^{*\mathrm{his}}+{\mathrm{\ϵ}}_t---\left(1\right)$

Wherein, represent the wind power prediction value of scheduling slot, for the wind power maximum output value of history scheduling slot, p represents the exponent number of forecast model, φ ₀for constant term; φ ₁, φ ₂, φ _pfor autoregressive coefficient, ε _tfor random disturbances amount;

The predicated error of wind power is the difference between actual wind power value of a certain moment and prediction, namely has:

$e_{i,k}=P_{i,k}^{\mathrm{real}}-P_{i,k}^{\mathrm{forecast}}---\left(2\right)$

Wherein e _i,kfor the predicated error of history i-th day kth scheduling slot, P _i,k ^realfor the actual maximum wind performance number of history i-th day kth scheduling slot, for history i-th day kth scheduling slot prediction wind-powered electricity generation maximum output value;

If Ω (j) represent t before the set of nearest n prediction error value, if the probability of each element is 1/n in set omega (j), then the empirical distribution function of Ω (j) be described below:

${\hat{F}}_k\left(\mathrm{\ξ}\right)=\frac{1}{N}\mathrm{Num}\{e_i\≤\mathrm{\ξ}|e_i\∈\mathrm{\Ω}\left(j\right),i=\mathrm{1,2}...,n\}---\left(3\right)$

Wherein: function Num is used for asking in set omega (j) number of elements meeting specified criteria, ξ is a stochastic variable of predicated error, represent that predicated error is less than the probability of ξ, n is the sum of history data set, Ω (j) represent t before the set of nearest n prediction error value, e _irepresent the predicated error of i-th scheduling slot;

Suppose for empirical distribution function inverse function, then a α probabilistic forecasting interval of actual value is:

$[P_{t|t+k}^{\mathrm{real}}+{\hat{G}}_t\left({\mathrm{\α}}_1\right),P_{t|t+k}^{\mathrm{real}}+{\hat{G}}_t\left({\mathrm{\α}}_2\right)]---\left(4\right)$

Wherein for the wind power actual value of a kth scheduling slot, α ₁lower limit for wind power probability interval: α ₁=(1-α)/2, α ₂higher limit for wind power probability interval: α ₂=1-(1-α)/2, for empirical distribution function inverse function;

2) determination of wind-powered electricity generation scheduling scheme

Receive wind-powered electricity generation to realize maximum-norm, its principle makes to dispatch the wind-powered electricity generation networked can utilize electrical network that space can be utilized not jeopardize again the safety of electrical network to greatest extent by generated output;

Consider that wind power prediction can exist larger predicated error, the actual value can not getting rid of wind power is greater than the situation appearance of predicted value, and this installed capacity that preliminary dispatching office will be caused to obtain exceedes the receiving limit of electrical network, jeopardizes the safe operation of electrical network;

For this reason, α probability interval is adopted to carry out the arrangement of wind-powered electricity generation scheduling scheme, to realize reducing the uncertain risk brought to operation of power networks of wind power prediction to greatest extent, so the k period dispatches the installed capacity of wind-driven power networked and is:

$C_k^{\mathrm{sp},\mathrm{pre}}=\frac{C_{\mathrm{Base}}^{*}}{{\stackrel{\‾}{P}}_{k.\mathrm{Day}\_Q+1}^{*\mathrm{forcast}}\left(\mathrm{\α}\right)}\×P_k^{\mathrm{space}}$

$\mathrm{st}.P_k^{\mathrm{space}}\≤C_k^{\mathrm{sp},\mathrm{pre}}\≤C_{\mathrm{\Σ},w}---\left(5\right)$

Wherein: C _k ^spacerepresent the scheduling networking capacity of wind-powered electricity generation; represent the higher limit of wind power α probability interval prediction, calculated by formula (4), C ^* _baserepresent the reference capacity for carrying out wind power prediction, P _k ^spacerepresent the utilized space of electrical network, in formula (5), constraints limits scheduling networking installed capacity of wind-driven power C _k ^{sp, pre}the electrical network being greater than scheduling slot can utilize space P _k ^spacebe less than intrasystem total installed capacity of wind-driven power C _{Σ, w};

3) energy storage system capacity configuration

Formula (5) is the source net coordinated scheduling scheme based on the prediction of wind-powered electricity generation probability interval, but for each scheduling day, adopt α probability interval forecast interval can not ensure that wind power prediction is worth 100% probability and drops in this interval, this can cause the wind-powered electricity generation of arranged scheduling scheme can generated output cross the safe operation limit of electrical network situation occur, jeopardize the safe operation of electrical network; Therefore, need to utilize energy-storage system to carry out charging to ensure the safe operation of electrical network; Corresponding energy-storage system charge and discharge control strategy is as follows:

$\left\{\begin{array}{c}{E}_k^{\mathrm{ch}\arg e}=E_0+\underset{i=1}{\overset{N}{\mathrm{\Σ}}}(P_{\mathrm{ref}}^i\×\mathrm{\ΔT}\×{\mathrm{\η}}_{\mathrm{ch}\arg e})\\ E_k^{\mathrm{disch}\arg e}=E_0+\underset{i=1}{\overset{N}{\mathrm{\Σ}}}(P_{\mathrm{ref}}^i\×\mathrm{\ΔT}/{\mathrm{\η}}_{\mathrm{disch}\arg e})\end{array}\right.---\left(6\right)$

Wherein E _k ^charge, E _k ^dischargee ₀be expressed as energy-storage system rechargeable energy, discharge energy and initial state-of-charge, N is the charge and discharge control number of times of each scheduling slot energy-storage system, and Δ T is the basic controlling cycle of energy-storage system, and it is relevant with sampling time interval; η _charge, η _dischargerepresent the charge and discharge efficiency of energy-storage system respectively, P ⁱ _reffor i moment energy-storage system charge power reference value, its computing formula is as follows:

$P_{\mathrm{ref}}^i=P_{\mathrm{real}}^i-P_k^{\mathrm{space}}---\left(7\right)$

Wherein P ⁱ _refrepresent the i-th moment energy-storage system charge power reference value, P ⁱ _realrepresent the actual value of the i-th moment wind power, P _k ^spacerepresent that the electrical network of scheduling slot can utilize space;

In order to ensure that the wind-powered electricity generation scheduling networking scale arranged in any scheduling slot is no more than the safe operation limit of electrical network, then energy storage system capacity need be equipped with by the maximum of scheduling slot each in dispatching cycle, and its computing formula is as follows:

$C_w=\max \left(\underset{n=1}{\overset{N}{\∪}}{\left\{E\right\}}_{k,\mathrm{day}\_N}\right)---\left(8\right)$

Wherein N is total scheduling day, C _wfor the energy storage system capacity of required outfit, { E} _{k, day_N}represent the rechargeable energy of energy-storage system at a N days kth scheduling slot.