CN104135028A - Unit distribution method combining black-start optimization model and FCB (Fast Cut Back) working condition - Google Patents

Abstract

The invention discloses a unit distribution method combining a black-start optimization model and a FCB (Fast Cut Back) working condition in the technical field of power system running and control. The unit distribution method comprises the following steps of: setting an optimal power flow target function; setting a certain thermal power generation in a system unit as a FCB unit; setting the FCB working condition according to the characteristic of the unit; setting an estimate time to solve a solution problem of a variable constraint set; converting a non-differentiable function into a differentiable target function form and a group of inequality constraint conditions by a mathematically equivalent method; and solving a feasible solution in a current solution domain based on a multi-center correction interior-point optimal power flow algorithm of the optimal center parameter; Compared with the prior art, the calculation result according to the unit distribution method is obviously helpful for recovery efficiency of the system.

Description

The unit points distributing method of the black startup Optimized model of a kind of combination and FCB operating mode

Technical field

The invention belongs to power system operation and control technology field, relate in particular to black startup model optimization algorithm and the FCB Transformation of Unit points distributing method of considering FCB technology quick recovery system after a kind of having a power failure on a large scale.

Background technology

So-called black startup, after referring to that whole system is because of fault stoppage in transit, system all has a power failure (do not get rid of isolated minor scale power net and still maintain operation), in complete " black " state, do not rely on other web help, the black startup power supply by having self-startup ability in system provides startup power supply to the target generator of non self starting, it is connected to the grid again, drive the generating set of non self starting, expansion system is gradually recovered scope, finally realizes restoring electricity of whole system.The primary goal startup stage of black should be to start maximum power supplys with the fastest speed, recovers strong as far as possible generating capacity.At present, in relevant references, conventionally with optimal load flow algorithm, optimize black start-up course, but at the power producing characteristics of often not considering generator of choosing of target function.The power producing characteristics of generator has directly limited the active power that generator can send within the set time, especially FCB (FCB, Fast Cut Back fast cut back) unit has the ability of quick adjustment power, make generator recover the identical spent asynchronism(-nization) of power, ignore this point and may cause result of calculation in real system, to be applied.The in the situation that of the parallel adjusting of each generator, should make all generators all in climbing, at utmost utilize the startup ability of generator, start as early as possible the whole network.Therefore, on mathematics, optimization aim should be: in the generator that final generator output allocative decision makes to have started, the longest that is compared to other scheme shortest time the required adjusting time, and the Mathematical Modeling of the whole network self-healing start-up course is actually one and usings maximum minimum problem as target function and with the optimization problem of the bundle of altering an agreement.And in the conventional solution of calculating at current optimal load flow, for the bundle maximum minimum problem of altering an agreement, there is no good Mathematical method.In order to solve, obtain black-start scheme, existing black startup model is used the target of approximately equivalent conventionally, reduces and solves difficulty, but can not realize well optimization aim.

Black startup power supply type mostly is Hydropower Unit and small-sized thermoelectric generator, the effects limit such as the type of power supply, capacity and geographical position the formulation of black-start scheme, previous black startup power configuration scheme is because limited too much needs hardly optimize to calculate.In recent years, fast cut back (FCB, Fast Cut Back) technology started to come into one's own.The generator with FCB function refers under accidental conditions, can and get rid of portion of external supply load with grid disconnection, with the generator of islet operation.Along with reaching its maturity of FCB technology, large-scale thermal power machine through FCB function modifications, there is starting capacity large and start the features such as speed is fast, the position of black startup power supply and the range of choice of capacity after system blackout have been expanded, black startup unit is layouted has had selectance more flexibly in capacity dimension and Spatial Dimension, becomes the strong point of fast quick-recovery after system large-area power-cuts.

Summary of the invention

The present invention takes into full account the physics regulating characteristics of generator, be provided with and take the shortest optimal load flow target function as target of system recovery time, propose to describe the minimum Optimized model of bundle maximum of altering an agreement of practical problem, and by Mathematical treatment cleverly by this problem equivalent be interior point method can separate conclude a contract or treaty bundle optimal power flow problems, overcome the difficult point that solves the bundle maximum minimum problem of altering an agreement in OPF (Optimal power flow tradition optimal load flow), proposed to take into account the power system blackstart optimization calculative strategy of generator output characteristic.In conjunction with this Optimized model, proposed to be with islet operation under FCB operating mode, accelerate the method for system recovering process, and propose based on this method that FCB Transformation of Unit is layouted.

Technical scheme of the present invention is that the unit points distributing method of the black startup Optimized model of a kind of combination and FCB operating mode, comprises the following steps:

Step 1: optimal load flow target function minf (x) is set, and it is target compared to other scheme shortest time that the longest that of required adjusting time in the generator having started take in the present invention.

Step 2: it is FCB unit that a certain fired power generating unit in system is set, and according to this machine unit characteristic, FCB operating mode is rationally set, and the present invention reverts to target with the most applicable auxiliary system operating mode is set.

Step 3: estimated time T is set _pre, according to generator output curve (accompanying drawing 1), determine the meritorious units limits P of starter-generator within this estimated time _min(T _pre) and P _max(T _pre).

Step 4: utilize the method for mathematics equivalence, by the target function of non-differentiability be converted into can be micro-target function form and one group of inequality constraints condition.

Step 5: the feasible solution in proofreading and correct with the multicenter based on optimum Center Parameter in the point current solution of optimal load flow Algorithm for Solving territory.If optimal load flow is not restrained, get back to step 3, in feasible region, amplify estimated time T _pre.If optimal load flow convergence, illustrates within the little estimated time and can find optimal solution, it must be the optimal solution in large solution space, this TV station generator recovery scheme has solved.

Step 6: select next to treat starter-generator, previous step is started to successful generator and add starter-generator group.

Step 7: repeating step 3 is to step 6, until all target generators are all restrained.If the optimal load flow of all target generators is not restrained, illustrate current system structure or configuration under cannot complete recovery control operation, need to change system configuration.If all generators start successfully, record FCB unit and transform herein the needed time of recovery system.

Step 8: repeating step 2 to 7, relatively FCB transformation is when diverse location, and the time that recovery system consumes, the fastest recuperator is the optimum position of layouting.Further,

In step 1, optimal load flow target function is made as to the adjusting shortest time of the longest generator of required adjusting time.Might as well establish and have m platform generator successfully to start, so, while starting m+1 platform generator, target function is:

$\min f\left(x\right)=\underset{k=1}{\overset{m}{\max }}\left\{t_k\right|\frac{P_k-P_0^k}{r_{\mathrm{pk}}}\left|\right\}---\left(1\right)$

Wherein, P _kbe k platform generator active power to be asked, P ₀ ^kfor current k platform generator active power, || represent to take absolute value, t _krepresent the adjusting time that k platform generator is required, r _pkthe climbing rate that represents k platform generator.Optimization aim is to make in all charged generators, the climbing minimal time of the generator that the required climbing time is the longest.

In step 2, FCB technology can only only have the thermal power plant in system can be set to FCB unit for fired power generating unit, and after transformation, the characterisitic parameters such as climbing rate of unit should be made modification according to actual conditions.Existing document considers FCB technology to be applied in system recovery, but does not provide the concrete operational mode of isolated island.Because black startup stage load recovers not yet comprehensively, the reactive power that circuit produces is often superfluous, when dropping into circuit, need to drop into a small amount of load for balance reactive power, also just cause generator need to send more active power, reduce system resume speed black startup stage.Assumed fault occurs in beyond FCB unit certain limit, can be with certain capacity stable operation after FCB load rejection, now should preferentially guarantee that circuit that susceptance value is large is in the state of putting into operation, in the scope allowing at unit output, guarantee that a certain amount of load makes trend convergence.When power-off, make isolated island that FCB is with maintain this state operation, the startup stage of can avoiding black, other unit of generating capacity deficiency starts some important line for a long time.

In step 3, as shown in generator output curve, R is generator starting capacity, and C is the generated power upper limit of exerting oneself, and D is that generator minimum technology is exerted oneself.T ₀represent that generator starts charging constantly, t ₁for the generator connecting in parallel with system moment, t ₂for reaching minimum technology, generator exerts oneself constantly.Reaching before minimum technology exerts oneself, generator can only upwards be climbed; Reaching after minimum technology exerts oneself, can upwards climb and also can climb downwards, but maximum can not surpass generator heap(ed) capacity, minimum can not be less than minimum technology exerts oneself, as shown by the arrows in Figure 1.R _pfor climbing rate, represent the speed that generated power is exerted oneself and regulated, namely power producing characteristics slope of a curve.

The inequality constraints of black startup model mainly comprises the constraint of node voltage bound, generated power is exerted oneself and idle units limits.As seen from Figure 1, there are relation power and the climbing time that due to generator, can promote, and therefore, the meritorious units limits of generator is one and becomes bound constraint:

$\left\{\begin{array}{c}{U}_{\min }\≤U\≤U_{\max }\\ P_{\min }\left(t\right)\≤P_G\≤P_{\max }\left(t\right)\\ Q_{\min }\≤Q_G\≤Q_{\max }\end{array}\right.$

Wherein, P _minand P (t) _max(t) the meritorious scope of exerting oneself of the m platform generator that expression has started, so active power constraint is the bundle of altering an agreement about time t.

From formula (2), generated power units limits be different from generally conclude a contract or treaty bundle OPF problem, solution space time to time change; And start-up time t span, depend on again the power shortage of system and the final method of salary distribution thereof, we wish the target of solving-optimizing.That is to say, in the bundle OPF model of altering an agreement, solve prioritization scheme and must first determine the scope retraining, and will determine that the scope of constraint first knows the result of prioritization scheme again.

The essence that solves of analyzing OPF model is to find optimal solution in the solution space consisting of variable bound collection.For change restricted problem is converted into and determines restricted model, the present invention is a prior given estimated time T before solving-optimizing problem _pre, determine the meritorious units limits P within this time period _min(T _pre) and P _max(T _pre), then the variable using time t as target function.The constriction of concludeing a contract or treaty that the constriction of altering an agreement is converted into after estimated time to variable range is set, if OPF does not restrain within the estimated time, in the situation that power regulates space in addition, amplifies estimated time value, increase solution space scope.The optimization aim proposing due to the present invention is the shortest time that makes each step starter-generator, if can find optimal solution within the little estimated time, it must be the optimal solution in large solution space.By this processing mode, be actually the constriction of concludeing a contract or treaty that the constriction of altering an agreement is converted into variable range.

In step 4, in order to represent that the span of variable t is subject to the restriction of active power, target function can be equivalent to one group of inequality constraints condition can micro-form:

minf(P _G,Q _G,e,f,t′)＝t′

$\begin{array}{cc}s.t.& t^{\′}\≥{\left(\frac{P_k{P}_{0k}}{r_{\mathrm{pk}}}\right)}^2\end{array}$

In step 8, system recovery time under the various schemes of layouting, not only relevant with the characterisitic parameter of FCB unit itself, also around transmit electricity environment etc. because have close ties with unit, therefore layouting of FCB unit is not simply large capacity unit to be transformed, the present invention's system from having a power failure on a large scale is recovered to consider, selects the fired power generating unit that after transformation, system is recovered the soonest to layout.

Effect of the present invention is, takes into full account generator physical characteristic, to recover model in the past, is optimized, and tries to achieve the shortest start-up time and the trend result that tally with the actual situation.This Optimized model is applied to FCB unit in conjunction with FCB operating mode layouts in the research of scheme.Compare with conventional method, result of calculation of the present invention has remarkable help to the recovery efficiency of system.

Accompanying drawing explanation

Fig. 1 generator output curve chart

Fig. 2 IEEE-30 node modular system

Generator climbing time and the system of Fig. 3 the present invention and conventional method are recovered total time comparison

Fig. 4 Guangdong Power Grid East Guangdong subregion rack schematic diagram

System recovering process under the complete black state of Fig. 5 East Guangdong subregion

Fig. 6 East Guangdong subregion Honghai Bay, Guangdong power plant band islet operation recovering process

Embodiment

Below in conjunction with accompanying drawing, the present invention is described in more detail.Should be emphasized that, following explanation is only exemplary, rather than in order to limit the scope of the invention and to apply.

The Optimized model that test the present invention proposes adopts IEEE-30 node modular system (as Fig. 2).Example comprises 30 nodes, 6 generators, 4 transformers and 37 transmission lines.Generator parameter is as shown in table 1.

Table 1 IEEE-30 node system generator parameter

Wherein node 1,2,5,8,11 and 13 is generator node, and corresponding generator numbering is respectively 1,2,3,4,5 and 6; Branch road 4-12,6-9,6-10 and 27-28 are transformer branch road, and other branch road is transmission line.No. 1 generator that is positioned at node 1 of take is black startup power supply, and node 2 No. 2 generators, No. 3 generators of node 5 are, No. 5 generators of No. 4 generators of node 8, node 11 and No. 6 generators of node 13 are target generator.Wherein transformer branch road start-up time is 10 minutes, and transmission line start-up time is 5 minutes.

Suppose that current system state is that No. 1, No. 2, No. 3 generators start, treat that starter-generator is No. 6 generators, power transmission path is 2-4-12-13, and it is 10 minutes that the estimated time is set, this step calculate before each active power and the power prediction bound of starter-generator is as shown in table 2.Wherein 2, No. 3 generators, owing to just starting, do not reach minimum technology and exert oneself, and are merely able to upwards climbing, and generator 1 has reached minimum technology and exerts oneself, and can upwards climb and also can climb downwards, but the minimum minimum technology that can not surpass exert oneself 0.3 _p.u..

The constraint of table 2 generator power

By target function be equivalent to one can micro-form target, and three inequality constraints conditions.In the multicenter correction of use based on optimum Center Parameter, current each of some optimal load flow Algorithm for Solving determined under constraints, the solution of optimal load flow, and when the estimated time equals 10 minutes, calculation of tidal current convergence, needn't expand solution space scope.Relatively minimum with the meritorious regulated quantity of exerting oneself of traditional Optimized model target, that is:

$\min f\left(x\right)=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{(P_G^k-P_{G0})}^2$

After No. 6 generators of startup that obtain, power adjustments situation and required climbing time are as shown in table 3.

Table 3 starts the power adjustments amount after target generator

In the middle of this step is calculated, except the starting capacity 0.018 of No. 6 generators, line loss active power is approximately 0.0004, can see, because the target of setting of the present invention is to make that the longest generator of climbing time the shortest compared to other scheme, therefore in fact the distribution of trend distributes according to generator climbing rate, makes every generator climbing time equal as far as possible.And conventional target is another kind of typical power division mode, be every generator mean allocation power, but this method is not considered the physical characteristic of generator, so the large unit of climbing rate need to wait for that the unit that climbing rate is little regulates, and increases total adjusting time.Obtain starting required climbing time of all target generators and system total recovery time as shown in Figure 3.

Because this example network is comparatively simple, the estimated time is set to all can meet and solve requirement for 10 minutes, and the present invention has also applied in the middle of the large-scale example of calculation shows of Guangdong Power Grid, has obtained good effect.In complex network, if the estimated time of 10 minutes cannot meet the demands, get back to step 2, increase the estimated time, restart the calculating of above-mentioned steps.The above, fully verified this system recovery and optimization model, compares with conventional method, and the present invention can significantly improve system and recover efficiency.

The impact on black start-up course in conjunction with Guangdong Power Grid example of calculation shows explanation FCB operating mode band islet operation again.Take east Guangdong Province as example, and all there are self-startup ability in Shantou Power Plant, Mei County power plant and Honghai Bay, Guangdong power plant, as shown in Figure 4.

Calculate respectively following two kinds of situations, the effect that checking FCB unit band islet operation recovers system:

(1) system is entirely black, and only three units with self-startup ability provide primary power, recovers the whole network.

(2) FCB of Honghai Bay, Guangdong power plant unit band islet operation, Mao Huzhan, Rongjiang station are charged, remain on low power operating state.

Because Honghai Bay, Guangdong power plant is the generating set belonging under 500KV rack, Honghai Bay, Guangdong factory---Mao Huzhan---Rongjiang station line admittance is very large, if system is entirely black under initial condition, starting this section of circuit will spend long time and just can make in zonule meritorious, reactive power equilibrium, the generating capacity of Honghai Bay, Guangdong power plant can not effectively be sent help system and recover, as shown in Figure 5, the charged region of red representative in figure, the region starting is being attempted in blue representative.

If while breaking down, Honghai Bay, Guangdong Power Plant has the ability of fast cut back band islet operation, after fault, Honghai Bay, Guangdong power plant can be with Mao Huzhan, a small amount of load in Rongjiang station to maintain running status, supports rapidly the system of Meizhou subregion and recovers, as shown in Figure 6 after recovering the power supply of Jing Hai B factory.From software, calculate demonstration and can see, FCB operating mode is applied in black start-up course, the initial stage recovery of system is had to great help.

The Optimized model proposing due to the present invention can solve the shortest startup scheme of the system time that is restored, and innovatively FCB operating mode is in a planned way incorporated in black start-up course, further quickening system resume speed, so the present invention has proposed again based on the shortest FCB Transformation of Unit points distributing method of system recovery time in conjunction with these two kinds of methods.East Guangdong subregion has 7 larger thermal power plants, is set to respectively FCB unit, according to line admittance value and the unit output upper limit, FCB operating mode is rationally set, and computing system recovers the required time, and result of calculation is as shown in table 4.Therefore because Honghai Bay, Guangdong power plant belongs to the main transmission of electricity rack of 500kV, if complete black startup will be more difficult, if there is FCB function to make the main transmission line can on-load islet operation, the effect that improves system resume speed be also the most obvious.

After the FCB of each thermal power plant of table 4 transformation, system is recovered required time (min)

This test macro is only preferably embodiment of the present invention; but protection scope of the present invention is not limited to this; anyly be familiar with those skilled in the art in the technical scope that the present invention discloses, the variation that can expect easily or replacement, within all should being encompassed in protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection range of claim.

Claims (5)

1. in conjunction with a unit points distributing method for black startup Optimized model and FCB operating mode, it is characterized in that comprising the following steps:

Step 1: optimal load flow target function minf (x) is set, and the longest that of required adjusting time in the generator having started of take is target compared to other scheme shortest time;

Step 2: it is FCB unit that a certain fired power generating unit in system is set, and according to this machine unit characteristic, FCB operating mode is set, and reverts to target operating mode is set with the most applicable auxiliary system;

Step 3: estimated time T is set _pre, according to generator output curve, determine within the described estimated time the meritorious units limits P of starter-generator _min(T _pre) and P _max(T _pre);

Step 4: utilize mathematics equivalent method, by the target function of non-differentiability be converted into can be micro-target function form and one group of inequality constraints condition;

Step 5: the feasible solution in proofreading and correct with the multicenter based on optimum Center Parameter in the point current solution of optimal load flow Algorithm for Solving territory, if optimal load flow is not restrained, get back to step 3, in feasible region, amplify estimated time T _pre; If optimal load flow convergence, it is the optimal solution in large solution space;

Step 6: select next to treat starter-generator, previous step is started to successful generator and add starter-generator group;

Step 7: repeating step 3 is to step 6, until all target generators are all restrained, if the optimal load flow of all target generators is not restrained, changes system configuration; If all generators start successfully, record this needed time of FCB Transformation of Unit recovery system;

Step 8: repeating step 2 is to step 7, relatively FCB transformation is when diverse location, and the time that recovery system consumes, selecting the fastest recovery time is the optimum position of layouting.

2. as claimed in claim 1 in conjunction with the unit points distributing method of black startup Optimized model and FCB operating mode, in wherein said step 1, optimal load flow target function is made as to the adjusting shortest time of the longest generator of required adjusting time, if there has been m platform generator successfully to start, while starting m+1 platform generator, optimal load flow target function is:

$\min f\left(x\right)=\underset{k=1}{\overset{m}{\max }}\left\{t_k\right|\frac{P_k-P_0^k}{r_{\mathrm{pk}}}\left|\right\}$

Wherein, P _kbe k platform generator active power to be asked, P ₀ ^kfor the active power of current k platform generator, || represent to take absolute value, t _krepresent the adjusting time that k platform generator is required, r _pkthe climbing rate that represents k platform generator, optimization aim is to make in all charged generators, the climbing minimal time of the generator that the required climbing time is the longest.

3. as claimed in claim 1 in conjunction with the unit points distributing method of black startup Optimized model and FCB operating mode, in wherein said step 2, the reduction of system resume speed startup stage that described FCB unit band islet operation being black to solve.

4. as claimed in claim 1 in conjunction with the unit points distributing method of black startup Optimized model and FCB operating mode, in wherein said step 3, generator can only upwards be climbed before exerting oneself reaching minimum technology; Reaching after minimum technology exerts oneself, can upwards climb and also can climb downwards, wherein, maximum can not surpass generator heap(ed) capacity, and minimum can not be less than minimum technology and exert oneself; The meritorious units limits of generator is one and becomes bound constraint:

$\left\{\begin{array}{c}{U}_{\min }\≤U\≤U_{\max }\\ P_{\min }\left(t\right)\≤P_G\≤P_{\max }\left(t\right)\\ Q_{\min }\≤Q_G\≤Q_{\max }\end{array}\right.$

Wherein, P _minand P (t) _max(t) the meritorious scope of exerting oneself of the m platform generator that expression has started;

The constriction of concludeing a contract or treaty that the constriction of altering an agreement is converted into after estimated time to variable range is set, if at estimated time T _preinterior OPF does not restrain, and in the situation that power regulates space in addition, amplifies estimated time value, increases solution space scope.

5. as claimed in claim 1 in conjunction with the unit points distributing method of black startup Optimized model and FCB operating mode, in wherein said step 4, described target function can be equivalent to one group of inequality constraints condition can micro-form:

minf(P _G,Q _G,e,f,t′)＝t′

$\begin{array}{cc}s.t.& t^{\′}\≥{\left(\frac{P_k{P}_{0k}}{r_{\mathrm{pk}}}\right)}^2\end{array}$

Wherein, the span of variable t is subject to the restriction of active power.