CN103401257A - Multi-source coordinated control method including wind power grid for coping with steep power slope at peak - Google Patents

Abstract

The invention provides a multi-source coordinated control method including a wind power grid for coping with a steep power slope at a peak. The method comprises the following steps: judging whether to start a control method or not according to front load and wind power prediction information; calculating the maximum adjustable margin of a BLR-AGC unit; calculating a steep slope rotation standby demand caused by a net load curve steep slope; establishing a load peak steep slope coordination strategy model to give different types of thermal power units and a resource scheduling power generation plan in different future time periods. The coordinated control method provided by the invention facilitates the power grid to go through peak dangerous time periods comprising steep slope events successfully, circuit overload, low frequency decreased load and even large-scale power failure accidents of the power grid, caused by poorer matching of a generator set, can be avoided effectively, and safe and stable running of the power grid is ensured.

Description

The multi-source coordination type control method that contains wind-powered electricity generation electrical network reply power abrupt slope, peak period

Technical field

The present invention relates to the dispatching of power netwoks solution formulation technology under large-scale wind power access background, be specifically related to a kind of multi-source coordination type control method that contains wind-powered electricity generation electrical network reply power abrupt slope, peak period.

Background technology

At present, large-scale wind power accesses the problem of bringing to electric power netting safe running and highlights all the more, some special operation is Japan-China, and wind power continues in peak period to reduce, and can make the relatively former load curve of peak period net load curve (load value checking electricity value) more precipitous.In case the peak period wind power falls suddenly, after with the rapid soaring effect of load, superposeing, larger variation will occur in interior net load in short-term, the abrupt slope of even breaking through the electrical network regulating power limit occur approaching.Abrupt slope is very big change of load curve shape not only, and also there is larger predicated error in itself.During generation abrupt slope, peak period event operating mode, electrical network must accurately calculate the abrupt slope spinning reserve demand that is caused by the abrupt slope predicated error.The abrupt slope prediction is the basis that meritorious coordination strategy is formulated in the electrical network peak period, and it predicts the outcome and has comparatively significantly time, amplitude and rate of change predicated error.Wherein, in certain period, except the amplitude prediction, itself there is certain error ME ₁(magnitude error, ME) outward, the time prediction error also can produce extra amplitude predicated error ME ₂.ME ₁With ME ₂Superposition value has formed amplitude predicated error ME that should the period jointly.ME is the principal element that affects electrical network spinning reserve demand, in the net load mathematics prediction under nominal situation, and ME ₁With ME ₂Amplitude all little, therefore the ME after both stacks is in BLR-AGC unit adjustable range, can all transfer to the BLR-AGC unit and dissolve, and also there is no need clearly to distinguish.Yet in the prediction of abrupt slope, ME ₁With ME ₂Amplitude may be larger, and the ME after stack probably exceeds BLR-AGC unit adjustable range, to electrical network, arranges spinning reserve to cause very big difficulty, also greatly increased electrical network simultaneously and formulated the difficulty of generating and alternative plan.If scheduling resource and control measures Shortcomings, electrical network probably can't be getted over abrupt slope smoothly, thereby produce power shortage, further bring out even major accident of low-frequency low-voltage load shedding.

Common high wind-powered electricity generation permeability dispatching of power netwoks scheme, lack the consideration to the abrupt slope spinning reserve demand that causes because of the abrupt slope predicated error at present, generally can only the adaptive peak operating mode that during phase, based model for load duration rises, wind-powered electricity generation continues slow decreasing.Under the background that installed capacity of wind-driven power constantly increases, by conventional method, formulate the dispatching of power netwoks scheme, in case wind power occurs in some special operation day, suddenly fall, follow generation net load curve to skyrocket, electrical network is emergent power vacancy event very easily, and safe operation is subject to serious threat.Therefore, under large-scale wind power access background, rationally utilize multiple scheduling resource, the accurate abrupt slope spinning reserve demand that is caused by abrupt slope time prediction error of calculating, and include these factors in novel meritorious scheduling strategy, to system reduce power shortage peak period, to get over dangerous working condition smoothly useful.

Summary of the invention

The object of the present invention is to provide a kind of multi-source coordination type control method that contains wind-powered electricity generation electrical network reply power abrupt slope, peak period.

The object of the present invention is achieved like this: method comprises the following steps:

Step 1: according to the place ahead load and wind-powered electricity generation prediction curve, judge whether control method starts; If the current on-line scheduling period is τ ₀, and obtained the place ahead, peak period N _preIndividual pre-scheduling period, N altogether _OnIndividual on-line scheduling period net load predicted value.To the place ahead on-line scheduling period τ _q, the corresponding pre-scheduling period is t _p, the load value that the BLO-AGC unit is followed the trail of

For:

$L_{\mathrm{on}}^{{\mathrm{\&tau;}}_q}=L_{\mathrm{net}}^{{\mathrm{\&tau;}}_q}-P_{\mathrm{BLR}}^{\mathrm{base}}-\underset{i\&Element;G_{\mathrm{Non}}}{\mathrm{\&Sigma;}}{\mathrm{<figure-callout\; id="1"\; label="P\; N\; ln"\; filenames="HSA0000093289900000021.png"\; state="\{\{state\}\}">P</figure-callout>}}_{N\ln }^{1\mathrm{plan},i}---\left(1\right)$

In formula:

For the whole basic point performance number of BLR-AGC unit, Peak climbing generally remains unchanged; Be that i platform Non-AGC unit (comprising the Non-AGC unit that can participate in online scheduling time level adjusting) is at pre-scheduling period t _pThe power planning value.

Two Rule of judgment that start the abrupt slope control strategy are as follows.

Condition 1: adjacent two on-line scheduling period net load fluctuations have surpassed the peak response ability of BLO-AGC unit.

$L_{\mathrm{on}}^{{\mathrm{\&tau;}}_q}-L_{\mathrm{on}}^{{\mathrm{\&tau;}}_{q-1}}>\mathrm{\&sigma;}\underset{j\&Element;G_{\mathrm{BLO}}}{\mathrm{\&Sigma;}}{r}_j\mathrm{\&Delta;\&tau;}---\left(2\right)$

In formula: σ is the nargin coefficient; r _jIt is j platform BLO-AGC unit regulations speed; Δ τ is on-line scheduling time stage time width; G _BLOFor the set of BLO-AGC unit.

Condition 2: through q on-line scheduling after the period net load accumulation amplification surpassed the peak response ability of BLO-AGC unit.

$L_{\mathrm{on}}^{{\mathrm{\&tau;}}_q}-L_{\mathrm{on}}^{{\mathrm{\&tau;}}_0}>\underset{j\&Element;G_{\mathrm{BLO}}}{\mathrm{\&Sigma;}}\min \{(P_j^{\max }-P_j^{{\mathrm{\&tau;}}_0}),q\&CenterDot;\mathrm{\&Delta;\&tau;}\&CenterDot;r_j\}---\left(3\right)$

If the place ahead on-line scheduling period τ _qMeet any one in formula (2) or formula (3) criterion, by τ _Q-1Period is to τ _qThe net load change of period is defined as abrupt slope, starts simultaneously the abrupt slope coordination strategy.

Step 2: calculate the maximum adjustable surplus amount of BLR-AGC unit; After the ultra-short Time level net load measured value of a plurality of operation day is added up with the deviation of on-line scheduling time stage net load predicted value, under high confidence level m, can obtain ultra-short Time level net load and predict that positive and negative deviation is respectively ε _{Net, (+)}With ε _{Net, (-)}.In certain on-line scheduling period, if exerting oneself, all scheduling resource plans just with the net load predicted value, equate with BLR-AGC unit basic point performance number sum, meet the requirement of on-line scheduling time stage power-balance, the BLR-AGC unit exists and is just regulating surplus S in the period at an on-line scheduling _BLRFor:

$S_{\mathrm{BLR}}=P_{\mathrm{BLR}}^{\max }-P_{\mathrm{BLR}}^{\mathrm{base}}-{\mathrm{\&epsiv;}}_{\mathrm{net},(+)}---\left(4\right)$

In formula:

For all BLR-AGC unit variable capacity upper limits.

Step 3: calculate abrupt slope positive rotation stand-by requirement; If n abrupt slope of net load curve occurs in τ _M-1Period is to τ _mPeriod, and amplitude is

τ _iPeriod (τ _i<τ _m) corresponding abrupt slope positive rotation stand-by requirement amount

For:

$R_{\mathrm{RRE},n}^{{\mathrm{\&tau;}}_i}=\left\{\begin{array}{c}\max \{0,D^{{\mathrm{\&tau;}}_m}-D^{{\mathrm{\&tau;}}_i}\},{\mathrm{\&tau;}}_m-{\mathrm{\&Delta;\&tau;}}_{\mathrm{error}}\&le;{\mathrm{\&tau;}}_i\&le;{\mathrm{\&tau;}}_{m-1}\\ 0,{\mathrm{\&tau;}}_i<{\mathrm{\&tau;}}_m-{\mathrm{\&Delta;\&tau;}}_{\mathrm{error}}\end{array}\right.---\left(5\right)$

The physical meaning of formula (5) is: if τ _iPeriod when generation corresponding to n abrupt slope in window, and should the period corresponding net load change amount less than This period will exist by n abrupt slope and may arrive in advance and the positive rotation stand-by requirement that produces; If net load change amount that should the period more than or equal to

Should the period by the positive rotation stand-by requirement that produces that arrives in advance of n abrupt slope, be zero.In addition, outside window online period when n abrupt slope occurred, the spinning reserve demand that this abrupt slope produces is also zero.

To the place ahead N _OnN in the individual on-line scheduling period _RAll calculate on individual abrupt slope

After, during due to the generation on these abrupt slopes, window may exist overlappingly, and certain online period may corresponding a plurality of positive rotation stand-by requirements like this.Should get the maximum in these positive rotation stand-by requirements this moment.Final like this place ahead τ that obtains _iThe abrupt slope positive rotation stand-by requirement of period is:

$R_{\mathrm{RRE}}^{{\mathrm{\&tau;}}_i}=\max \left\{R_{\mathrm{RRE},n}^{{\mathrm{\&tau;}}_i}\right\}(n=1....N_R)---\left(6\right)$

Step 4: set up abrupt slope, peak period control strategy model; The simulated target function is as follows:

$\min \left\{\underset{\mathrm{\&tau;}=1}{\overset{N_{\mathrm{on}}}{\mathrm{\&Sigma;}}}\right[f_{\mathrm{BLO}}^{\mathrm{\&tau;}}+f_{\mathrm{IL}}^{\mathrm{\&tau;}}+f_{\mathrm{eps}}^{\mathrm{\&tau;}}+f_{\mathrm{lack}}^{\mathrm{\&tau;}}+f_{\mathrm{LS}}^{\mathrm{\&tau;}}]+\underset{t=1}{\overset{N_{\mathrm{pre}}}{\mathrm{\&Sigma;}}}{f}_{\mathrm{Non}1}^t\}$

$\left\{\begin{array}{c}{f}_{\mathrm{BLO}}^{\mathrm{\&tau;}}=\underset{{j\&Element;G}_{\mathrm{BLO}}}{\mathrm{\&Sigma;}}(k_{\mathrm{BLO},j}{P}_{\mathrm{BLO},j}^{\mathrm{\&tau;}}+u_{\mathrm{BLO},j}{R}_{\mathrm{BLO},j}^{\mathrm{\&tau;}})\\ f_{\mathrm{IL}}^{\mathrm{\&tau;}}=k_{\mathrm{IL}}{P}_{\mathrm{IL}}^{\mathrm{\&tau;}}\\ f_{\mathrm{eps}}^{\mathrm{\&tau;}}=k_{\mathrm{eps}}{P}_{\mathrm{eps}}^{\mathrm{\&tau;}}+u_{\mathrm{eps}}{R}_{\mathrm{eps}}^{\mathrm{\&tau;}}\\ f_{\mathrm{lack}}^{\mathrm{\&tau;}}=k_{\mathrm{lack}}{P}_{\mathrm{lack}}^{\mathrm{\&tau;}}\\ f_{\mathrm{LS}}^{\mathrm{\&tau;}}=k_{\mathrm{LS}}\mathrm{\&Delta;}{L}^{\mathrm{\&tau;}}\\ f_{\mathrm{Non}1}^t=k_{\mathrm{Non}}\mathrm{\&Delta;}{P}_{\mathrm{Non}1}^t\end{array}---\left(7\right)\right.$

In formula: N _OnHop count during for predictable all on-line schedulings of net load; N _preHop count during for predictable all pre-schedulings of net load; For all BLO-AGC unit operation costs of τ period,

For τ period j platform BLO-AGC unit power output,

For positive rotation reserve level, k _{BLO, j}, u _{BLO, j}For unit cost of electricity-generating and spinning reserve cost;

For τ period interruptible load cost, For τ period interruptible load uses total amount, k _ILFor the interruptible load average unit cost;

For urgent peaking power source operating cost of τ period,

For urgent peaking power source of τ period uses total amount,

For positive rotation reserve level, k _eps, u _epsCorresponding unit cost of electricity-generating and unit spinning reserve cost;

For the adjustable surplus of τ period BLR-AGC unit in step 2 weakens cost,

For power shortage amount, k _LackFor the power shortage unit cost; For τ period limited load cost, Δ L ^τFor τ period limited load total amount, k _LSFor unit limited load average unit cost.

During for pre-scheduling period t, only participate in the Non-AGC unit (G of pre-scheduling time stage regulation and control _Non1Set) total adjustment cost,

For G _Non1Set aggregate capacity adjustment amount, k _Non1For G _Non1The set unit destroys the average adjusted cost of the original plan.Formula (7) model meets following constraint:

The pre-scheduling period is t _p, the on-line scheduling period is τ _qThe time power-balance be constrained to:

$\begin{array}{c}\underset{i\&Element;G_{\mathrm{Non}1}}{\mathrm{\&Sigma;}}{P}_{\mathrm{Non}1\mathrm{plan},i}^{t_p}+{\mathrm{\&Delta;P}}_{\mathrm{Non}1}^{t_p}+\underset{j\&Element;G_{\mathrm{BLO}}}{\mathrm{\&Sigma;}}{P}_{\mathrm{BLO},j}^{{\mathrm{\&tau;}}_q}+P_{\mathrm{eps}}^{{\mathrm{\&tau;}}_q}\\ +P_{\mathrm{BLR}}^{\mathrm{base}}+L_{\mathrm{net}}^{{\mathrm{\&tau;}}_q}-P_{\mathrm{lack}}^{{\mathrm{\&tau;}}_q}-{\mathrm{\&Delta;L}}^{{\mathrm{\&tau;}}_q}\end{array}---\left(8\right)$

In formula:

For t _pPeriod G _Non1In set, unit is worth in the original plan, is known quantity.

The constraint of on-line scheduling time stage power bound:

$\left\{\begin{array}{c}{P}_{\mathrm{BLO},j}^{\min }\&le;P_{\mathrm{BLO},j}^{\mathrm{\&tau;}}\&le;P_{\mathrm{BLO},j}^{\max }\\ P_{\mathrm{eps}}^{\min }\&le;P_{\mathrm{eps}}^{\mathrm{\&tau;}}\&le;P_{\mathrm{eps}}^{\max }\end{array}\right.---\left(9\right)$

In formula:

For the upper and lower limit of BLO-AGC unit power output; Be respectively the upper and lower limit of Emergency Power power output.

On-line scheduling time stage unit ramping rate constraints:

${-r}_j\mathrm{\&Delta;\&tau;}\&le;P_{\mathrm{BLO},j}^{\mathrm{\&tau;}}-P_{\mathrm{BLO},j}^{\mathrm{\&tau;}-1}\&le;r_j\mathrm{\&Delta;\&tau;}---\left(10\right)$

In formula: r _jFor BLO-AGC unit regulations speed.

The BLR-AGC unit is regulated the nargin constraint:

$0\&le;P_{\mathrm{lack}}^{\mathrm{\&tau;}}\&le;S_{\mathrm{BLR}}---\left(11\right)$

The positive rotation Reserve Constraint:

$\left\{\begin{array}{c}\underset{j\&Element;G_{\mathrm{BLO}}}{\mathrm{\&Sigma;}}{R}_{\mathrm{BLO},j}^{\mathrm{\&tau;}}+R_{\mathrm{eps}}^{\mathrm{\&tau;}}+P_{\mathrm{IL}}^{\mathrm{\&tau;}}\&GreaterEqual;R_{\mathrm{RRE}}^{\mathrm{\&tau;}}\\ P_{\mathrm{BLO},j}^{\mathrm{\&tau;}}+R_{\mathrm{BLO},j}^{\mathrm{\&tau;}}\&le;P_{\mathrm{BLO},\mathrm{<figure-callout\; id="0"\; label="j\; max"\; filenames="HSA0000093289900000011.png"\; state="\{\{state\}\}">j</figure-callout>}}^{\max }\\ 0\&le;R_{\mathrm{BLO},j}^{\mathrm{\&tau;}}\&le;r_j\mathrm{\&Delta;\&tau;}\\ P_{\mathrm{eps}}^{\mathrm{\&tau;}}+R_{\mathrm{eps}}^{\mathrm{\&tau;}}\&le;P_{\mathrm{eps}}^{\max }\\ 0\&le;P_{\mathrm{IL}}^{\mathrm{\&tau;}}\&le;P_{\mathrm{IL}}^{\max }\end{array}\right.---\left(12\right)$

In formula:

For interruptible load uses the total amount upper limit.

Time dimension and the optimized variable of above model are all few, can adopt classical linear programming algorithm to solve, and the rolling that is applicable to control strategy is upgraded.In optimized variable, the interruptible load use amount

Limited load amount Δ L ^τWith Non-AGC unit overall adjustment amount

All with the total amount form, provide.If obtain

With Δ L ^τUntrivialo solution, by the load significance level of each interruptible load use cost from different nodes, provide interruptible load operational version and each node plan limited load amount.

Give G _Non1In set, the Non-AGC unit output is optimized submodel.

G _Non1The set unit output is optimized submodel:

$\min \underset{t=1}{\overset{N_{\mathrm{pre}}}{\mathrm{\&Sigma;}}}\underset{i\&Element;G_{\mathrm{Non}1}}{\mathrm{\&Sigma;}}{k}_{\mathrm{Non}1,i}{\mathrm{\&Delta;P}}_{\mathrm{Non}1,i}^t$

$\left\{\begin{array}{c}\underset{i\&Element;G_{\mathrm{Non}1}}{\mathrm{\&Sigma;}}{\mathrm{\&Delta;P}}_{\mathrm{Non}1,i}^t={\mathrm{\&Delta;P}}_{\mathrm{Non}1}^t\\ {\mathrm{\&Delta;P}}_{\mathrm{Non}1}^t\&GreaterEqual;0\\ \underset{\&OverBar;}{h}\&le;h(P_{\mathrm{Non}1}^t+{\mathrm{\&Delta;P}}_{\mathrm{Non}1}^t)\&le;\stackrel{\&OverBar;}{h}\end{array}\right.---\left(13\right)$

In formula:

For G _Non1In set, i platform Non-AGC unit is at the power adjustment of pre-scheduling period t, k _{Non1, i}For corresponding unit adjustment cost; H is G _Non1The physical constraint of set unit and network congestion constraint.Attention is solving G _Non1Network congestion when constraint of set unit, should using by BLO-AGC power of the assembling unit planned value, Emergency Power power planning value and each node plan limited load amount in the multi-period multi-source Coordination Model in abrupt slope as network known power input source.

Beneficial effect of the present invention is as follows:

(1) while being applicable to high wind-powered electricity generation permeability electrical network net load prediction curve and abrupt slope occurs, the formulation of electric network active scheduling strategy, this coordination strategy has accurately calculated the abrupt slope spinning reserve demand that is caused by abrupt slope time prediction error, rationally utilized number of different types scheduling resource in electrical network, be of value to when electrical network abrupt slope occurs in the net load peak period and get over dangerous periods.

(2) it is in service that the generation schedule that relies on this control strategy to formulate is put into the electrical network actual schedule, can effectively avoid because generating set coordinates poor electrical network generation circuit overload, the even massive blackout accident of UFLS of causing, favourable to guaranteeing power network safety operation.

(3) the cutting load amount that provides according to control strategy, dispatching of power netwoks department can hold in advance climbs the dangerous situation in peak, informs that in advance non-important load user carries out the restriction load and prepares, and reduces the loss.

The accompanying drawing explanation

Fig. 1 is time, amplitude and the rate of change predicated error schematic diagram of abrupt slope prediction;

Fig. 2 is control flow chart of the present invention;

Fig. 3 is net load predicted value figure.

Embodiment

The invention will be further described for example below in conjunction with accompanying drawing.

Embodiment 1: get certain province's actual electric network year load data this paper control strategy is analyzed.The Non-AGC unit that this electrical network has 4 BLR-AGC units, 15 can only participate in the pre-scheduling time stage to regulate, 1 can participate in online scheduling time level Non-AGC unit (being numbered G1), 3 identical BLO-AGC units of physical characteristic (equivalence is to be numbered G2 after 1 unit) and 1 urgent peaking power source (being numbered G3) of regulating; The wind-powered electricity generation total installation of generating capacity is 1260MW; The unit partial parameters is in Table 1, and the parameters such as operation of power networks situation are in Table 2.Following 2 pre-scheduling periods, totally 6 on-line scheduling period net load predicted values are shown in Pat1 curve in Fig. 3.Current being in of electrical network climbed the peak later stage, and fired power generating unit integral body is exerted oneself and is in a high position.

Table 1 unit parameter

?	G1	G2	G3	Interruptible load
					Maximum size (MW)	800	600	40	10
Lower bound of capacity (MW)	460	300	10	0
					Power initial value (MW)	490	480	--	--
Regulations speed (MW/min)	6	16	--	--
					K (unit/MW)	1400	300	3000	100
U (unit/MW)	1000	150	1600	--

Table 2 electrical network parameter

Parameter	Numerical value	Parameter	Numerical value
				k lack1(unit/MW/5min)	6000	S BLR(MW)	20
k LS(unit/MW/5min)	30000	Δτ error	2
				β(MW/Hz)	-1200	?	?

According to the place ahead period net load prediction case, find that period 0-1, period 1-2, period 2-3, the fluctuation of period 5-6 net load meet the entry condition 1 of abrupt slope control strategy, period 3-4 and the fluctuation of period 4-5 net load meet the entry condition 2 of control strategy, net load change is violent, the abrupt slope control strategy starts immediately, obtains unit generation plan and alternative plan in Table 3.

Unit generation and the alternative plan of table 3Pat1

Visible by table 3 result, G1 and G2 have reached the regulations speed limit or the unit capacity upper limit in a plurality of periods.Abrupt slope spinning reserve demand reached maximum 2 o'clock periods, was 68MW.This is mainly that time prediction error by period 2-3 causes, period 2-3 net load change reaches maximum 153MW.When the pre-scheduling period started, the Non-AGC unit was by changing former generation schedule, reduced G1 and G2 in the period 1 force level that goes out with the period 4, make G1 and G2 can be in getting over the abrupt slope process all one's effort Pan Feng.At the most violent period 2-3 of net load change, when G3 reached peak power output 40MW, the BLR-AGC unit had used the adjusting surplus of 13MW, had avoided the cutting load risk and had getted over smoothly all abrupt slopes.

Embodiment 2: identical with all parameters of electrical network in embodiment 1, only net load change is more violent, in 2-3 period and 3-4 period, has increased 10MW in respectively than embodiment 1, sees Pat2 curve in Fig. 3.The Pat2 generation schedule is in Table 4.

Unit generation and the alternative plan of table 4Pat2

With Pat1, compare, Pat2 is affected by more violent net load change, 2 o'clock periods, will meet the abrupt slope spinning reserve demand up to 78MW, and this forces G2 to compare Pat1 in the period 2 will prepare more spinning reserve.In addition, the period 3, after the adjusting surplus of using up urgent peaking power source and BLR-AGC unit, must just can get over this period by limited load 3MW smoothly.Visible from the Pat2 result, rely on multiple scheduling resource and control measures cooperation, although electrical network has been transferred multiple expensive scheduling resource, and sacrificed the adjusting surplus of BLR-AGC unit, but at utmost reduced electrical network limited load amount, at utmost reduce overall loss guaranteeing that electrical network is under the safety operation level prerequisite, had obvious economic benefit.

Claims (1)

1. contain the multi-source coordination type control method on wind-powered electricity generation electrical network reply power abrupt slope, peak period, it is characterized in that, method is as follows:

Step 1: according to the place ahead load and wind-powered electricity generation prediction curve, judge whether control method starts; The current on-line scheduling period is τ ₀, and obtained the place ahead, peak period N _preIndividual pre-scheduling period, N altogether _OnIndividual on-line scheduling period net load predicted value, to the place ahead on-line scheduling period τ _q, the corresponding pre-scheduling period is t _p, the load value that the BLO-AGC unit is followed the trail of

For:

$L_{\mathrm{on}}^{{\mathrm{\&tau;}}_q}=L_{\mathrm{net}}^{{\mathrm{\&tau;}}_q}-P_{\mathrm{BLR}}^{\mathrm{base}}-\underset{{i\&Element;G}_{\mathrm{Non}}}{\mathrm{\&Sigma;}}{P}_{\mathrm{Non}1\mathrm{plan},i---\left(1\right)}^{t_p}$

In formula:

For the whole basic point performance number of BLR-AGC unit, Peak climbing remains unchanged; Be that i platform Non-AGC unit is at pre-scheduling period t _pThe power planning value;

Two Rule of judgment that start the abrupt slope control method are as follows:

Condition 1: adjacent two on-line scheduling period net load fluctuations have surpassed the peak response ability of BLO-AGC unit,

$L_{\mathrm{on}}^{{\mathrm{\&tau;}}_q}-L_{\mathrm{on}}^{{\mathrm{\&tau;}}_{q-1}}>\mathrm{\&sigma;}\underset{{j\&Element;G}_{\mathrm{BLO}}}{\mathrm{\&Sigma;}}{r}_j\mathrm{\&Delta;\&tau;}---\left(2\right)$

In formula: σ is the nargin coefficient; r _jIt is j platform BLO-AGC unit regulations speed; Δ τ is on-line scheduling time stage time width; G _BLOFor the set of BLO-AGC unit;

Condition 2: through q on-line scheduling after the period net load accumulation amplification surpassed the peak response ability of BLO-AGC unit,

$L_{\mathrm{on}}^{{\mathrm{\&tau;}}_q}-L_{\mathrm{on}}^{{\mathrm{\&tau;}}_0}>\underset{{j\&Element;G}_{\mathrm{BLO}}}{\mathrm{\&Sigma;}}\min \{(P_j^{\max }-P_j^{{\mathrm{\&tau;}}_0}),q\&CenterDot;\mathrm{\&Delta;\&tau;}\&CenterDot;r_j\}---\left(3\right)$

The place ahead on-line scheduling period τ _qMeet any one in formula (2) or formula (3) criterion, by τ _Q-1Period is to τ _qThe net load change of period is defined as abrupt slope, starts simultaneously control method;

Step 2: calculate the maximum adjustable surplus amount of BLR-AGC unit; After the ultra-short Time level net load measured value of a plurality of operation day is added up with the deviation of on-line scheduling time stage net load predicted value, under high confidence level m, obtain ultra-short Time level net load and predict that positive and negative deviation is respectively ε _{Net, (+)}With ε _{Net, (-)}In certain on-line scheduling period, all scheduling resource plans are exerted oneself and just with the net load predicted value, are equated with BLR-AGC unit basic point performance number sum, meet the requirement of on-line scheduling time stage power-balance, the BLR-AGC unit exists and is just regulating surplus S in the period at an on-line scheduling _BLRFor:

$S_{\mathrm{BLR}}=P_{\mathrm{BLR}}^{\max }-P_{\mathrm{BLR}}^{\mathrm{base}}-{\mathrm{\&epsiv;}}_{\mathrm{net},(+)}---\left(4\right)$

In formula:

For all BLR-AGC unit variable capacity upper limits;

Step 3: calculate abrupt slope positive rotation stand-by requirement; N abrupt slope of net load curve occurs in τ _M-1Period is to τ _mPeriod, and amplitude is

τ _iPeriod (τ _i<τ _m) corresponding abrupt slope positive rotation stand-by requirement amount

For:

$R_{\mathrm{RRE},n}^{{\mathrm{\&tau;}}_i}=\left\{\begin{array}{c}\max \{0,D^{{\mathrm{\&tau;}}_m}-D^{{\mathrm{\&tau;}}_i}\},{\mathrm{\&tau;}}_m-{\mathrm{\&Delta;\&tau;}}_{\mathrm{error}}\&le;{\mathrm{\&tau;}}_i\&le;{\mathrm{\&tau;}}_{m-1}\\ 0,{\mathrm{\&tau;}}_i<{\mathrm{\&tau;}}_m-{\mathrm{\&Delta;\&tau;}}_{\mathrm{error}}\end{array}\right.---\left(5\right)$

The physical meaning of formula (5) is: if τ _iPeriod when generation corresponding to n abrupt slope in window, and should the period corresponding net load change amount less than

This period will exist by n abrupt slope and may arrive in advance and the positive rotation stand-by requirement that produces; If net load change amount that should the period more than or equal to

Should the period by the positive rotation stand-by requirement that produces that arrives in advance of n abrupt slope, be zero; In addition, outside window online period when n abrupt slope occurred, the spinning reserve demand that this abrupt slope produces is also zero;

To the place ahead N _OnN in the individual on-line scheduling period _RAll calculate on individual abrupt slope

After, should get the maximum in these positive rotation stand-by requirements, final like this place ahead τ that obtains this moment _iThe abrupt slope positive rotation stand-by requirement of period is:

$R_{\mathrm{RRE}}^{{\mathrm{\&tau;}}_i}=\max \left\{R_{\mathrm{RRE},n}^{{\mathrm{\&tau;}}_i}\right\}(n=1....N_R)---\left(6\right)$

Step 4: set up abrupt slope, peak period control strategy model; The simulated target function is as follows:

$\min \left\{\underset{\mathrm{\&tau;}=1}{\overset{N_{\mathrm{on}}}{\mathrm{\&Sigma;}}}\right[f_{\mathrm{BLO}}^{\mathrm{\&tau;}}+f_{\mathrm{IL}}^{\mathrm{\&tau;}}+f_{\mathrm{eps}}^{\mathrm{\&tau;}}+f_{\mathrm{lack}}^{\mathrm{\&tau;}}+f_{\mathrm{LS}}^{\mathrm{\&tau;}}]+\underset{t=1}{\overset{N_{\mathrm{pre}}}{\mathrm{\&Sigma;}}}{f}_{\mathrm{Non}1}^t\}$

$\left\{\begin{array}{c}{f}_{\mathrm{BLO}}^{\mathrm{\&tau;}}=\underset{{j\&Element;G}_{\mathrm{BLO}}}{\mathrm{\&Sigma;}}(k_{\mathrm{BLO},j}{P}_{\mathrm{BLO},j}^{\mathrm{\&tau;}}+u_{\mathrm{BLO},j}{R}_{\mathrm{BLO},j}^{\mathrm{\&tau;}})\\ f_{\mathrm{IL}}^{\mathrm{\&tau;}}=k_{\mathrm{IL}}{P}_{\mathrm{IL}}^{\mathrm{\&tau;}}\\ f_{\mathrm{eps}}^{\mathrm{\&tau;}}=k_{\mathrm{eps}}{P}_{\mathrm{eps}}^{\mathrm{\&tau;}}+u_{\mathrm{eps}}{R}_{\mathrm{eps}}^{\mathrm{\&tau;}}\\ f_{\mathrm{lack}}^{\mathrm{\&tau;}}=k_{\mathrm{lack}}{P}_{\mathrm{lack}}^{\mathrm{\&tau;}}\\ f_{\mathrm{LS}}^{\mathrm{\&tau;}}=k_{\mathrm{LS}}\mathrm{\&Delta;}{L}^{\mathrm{\&tau;}}\\ f_{\mathrm{Non}1}^t=k_{\mathrm{Non}}\mathrm{\&Delta;}{P}_{\mathrm{Non}1}^t\end{array}---\left(7\right)\right.$

In formula: N _OnHop count during for predictable all on-line schedulings of net load; N _preHop count during for predictable all pre-schedulings of net load;

For all BLO-AGC unit operation costs of τ period, For τ period j platform BLO-AGC unit power output,

For positive rotation reserve level, k _{BLO, j}, u _{BLO, j}For unit cost of electricity-generating and spinning reserve cost;

For τ period interruptible load cost,

For τ period interruptible load uses total amount, k _ILFor the interruptible load average unit cost;

For urgent peaking power source operating cost of τ period,

For urgent peaking power source of τ period uses total amount,

For positive rotation reserve level, k _eps, u _epsCorresponding unit cost of electricity-generating and unit spinning reserve cost;

For the adjustable surplus of τ period BLR-AGC unit in step 2 weakens cost,

For power shortage amount, k _LackFor the power shortage unit cost;

For τ period limited load cost, Δ L ^τFor τ period limited load total amount, k _LSFor unit limited load average unit cost.

During for pre-scheduling period t, only participate in the Non-AGC unit (G of pre-scheduling time stage regulation and control _Non1Set) total adjustment cost,

For G _Non1Set aggregate capacity adjustment amount, k _Non1For G _Non1The set unit destroys the average adjusted cost of the original plan; Formula (7) model meets following constraint:

The pre-scheduling period is t _p, the on-line scheduling period is τ _qThe time power-balance be constrained to:

$\begin{array}{c}\underset{i\&Element;G_{\mathrm{Non}1}}{\mathrm{\&Sigma;}}{P}_{\mathrm{Non}1\mathrm{plan},i}^{t_p}+{\mathrm{\&Delta;P}}_{\mathrm{Non}1}^{t_p}+\underset{j\&Element;G_{\mathrm{BLO}}}{\mathrm{\&Sigma;}}{P}_{\mathrm{BLO},j}^{{\mathrm{\&tau;}}_q}+P_{\mathrm{eps}}^{{\mathrm{\&tau;}}_q}\\ +P_{\mathrm{BLR}}^{\mathrm{base}}+L_{\mathrm{net}}^{{\mathrm{\&tau;}}_q}-P_{\mathrm{lack}}^{{\mathrm{\&tau;}}_q}-{\mathrm{\&Delta;L}}^{{\mathrm{\&tau;}}_q}\end{array}---\left(8\right)$

In formula: For t _pPeriod G _Non1In set, unit is worth in the original plan, is known quantity;

The constraint of on-line scheduling time stage power bound:

$\left\{\begin{array}{c}{P}_{\mathrm{BLO},j}^{\min }\&le;P_{\mathrm{BLO},j}^{\mathrm{\&tau;}}\&le;P_{\mathrm{BLO},j}^{\max }\\ P_{\mathrm{eps}}^{\min }\&le;P_{\mathrm{eps}}^{\mathrm{\&tau;}}\&le;P_{\mathrm{eps}}^{\max }\end{array}\right.---\left(9\right)$

In formula:

For the upper and lower limit of BLO-AGC unit power output;

Be respectively the upper and lower limit of Emergency Power power output;

On-line scheduling time stage unit ramping rate constraints:

${-r}_j\mathrm{\&Delta;\&tau;}\&le;P_{\mathrm{BLO},j}^{\mathrm{\&tau;}}-P_{\mathrm{BLO},j}^{\mathrm{\&tau;}-1}\&le;r_j\mathrm{\&Delta;\&tau;}---\left(10\right)$

In formula: r _jFor BLO-AGC unit regulations speed;

The BLR-AGC unit is regulated the nargin constraint:

$0\&le;P_{\mathrm{lack}}^{\mathrm{\&tau;}}\&le;S_{\mathrm{BLR}}---\left(11\right)$

The positive rotation Reserve Constraint:

$\left\{\begin{array}{c}\underset{j\&Element;G_{\mathrm{BLO}}}{\mathrm{\&Sigma;}}{R}_{\mathrm{BLO},j}^{\mathrm{\&tau;}}+R_{\mathrm{eps}}^{\mathrm{\&tau;}}+P_{\mathrm{IL}}^{\mathrm{\&tau;}}\&GreaterEqual;R_{\mathrm{RRE}}^{\mathrm{\&tau;}}\\ P_{\mathrm{BLO},j}^{\mathrm{\&tau;}}+R_{\mathrm{BLO},j}^{\mathrm{\&tau;}}\&le;P_{\mathrm{BLO},j}^{\max }\\ 0\&le;R_{\mathrm{BLO},j}^{\mathrm{\&tau;}}\&le;r_j\mathrm{\&Delta;\&tau;}\\ P_{\mathrm{eps}}^{\mathrm{\&tau;}}+R_{\mathrm{eps}}^{\mathrm{\&tau;}}\&le;P_{\mathrm{eps}}^{\max }\\ 0\&le;P_{\mathrm{IL}}^{\mathrm{\&tau;}}\&le;P_{\mathrm{IL}}^{\max }\end{array}\right.---\left(12\right)$

In formula:

For interruptible load uses the total amount upper limit;

The time dimension of above model and optimized variable adopt classical linear programming algorithm to solve, in optimized variable, and the interruptible load use amount

Limited load amount Δ L ^τWith Non-AGC unit overall adjustment amount

All with the total amount form, provide; If obtain

With Δ L ^τUntrivialo solution, by the load significance level of each interruptible load use cost from different nodes, provide interruptible load operational version and each node plan limited load amount;

Give G _Non1In set, the Non-AGC unit output is optimized submodel;

G _Non1The set unit output is optimized submodel:

$\min \underset{t=1}{\overset{N_{\mathrm{pre}}}{\mathrm{\&Sigma;}}}\underset{i\&Element;G_{\mathrm{Non}1}}{\mathrm{\&Sigma;}}{k}_{\mathrm{Non}1,i}{\mathrm{\&Delta;P}}_{\mathrm{Non}1,i}^t$

$\left\{\begin{array}{c}\underset{i\&Element;G_{\mathrm{Non}1}}{\mathrm{\&Sigma;}}{\mathrm{\&Delta;P}}_{\mathrm{Non}1,i}^t={\mathrm{\&Delta;P}}_{\mathrm{Non}1}^t\\ {\mathrm{\&Delta;P}}_{\mathrm{Non}1}^t\&GreaterEqual;0\\ \underset{\&OverBar;}{h}\&le;h(P_{\mathrm{Non}1}^t+{\mathrm{\&Delta;P}}_{\mathrm{Non}1}^t)\&le;\stackrel{\&OverBar;}{h}\end{array}\right.---\left(13\right)$

In formula:

For G _Non1In set, i platform Non-AGC unit is at the power adjustment of pre-scheduling period t, k _{Non1, i}For corresponding unit adjustment cost; H is G _Non1The physical constraint of set unit and network congestion constraint; Solving G _Non1Network congestion when constraint of set unit, should using by BLO-AGC power of the assembling unit planned value, Emergency Power power planning value and each node plan limited load amount in the multi-period multi-source Coordination Model in abrupt slope as network known power input source.

Images