CN104466953A - Power grid cascading failure restraining method - Google Patents

Abstract

The invention discloses a power grid cascading failure restraining method and belongs to the technical field of safety protection of power systems. According to the method, a node which has largest impact on an overload branch circuit is worked out, the generator tripping and load shedding amount range is determined and optimized many times, generator tripping and load shedding operations are carried out on a power grid network according to the generator tripping and load shedding amount optimized every time, and whether the current value of the overload branch circuit at the next time is within the safety range or not is detected. The method can continuously and cyclically carry out generator tripping and load shedding amount optimization and load shedding operation until the current value of the overload branch circuit reaches the safety range or the number of times of optimization of the generator tripping and load shedding amount exceeds the maximum number. The method can optimize the generator tripping and load shedding amount when the branch circuit is overloaded, reduce loss and effectively restrain cascading failure at the same time.

Description

A kind of suppressing method of power grid cascading fault

Technical field

The invention belongs to power system security guard technology field, be specifically related to a kind of suppressing method of power grid cascading fault.

Background technology

Bulk power grid faces many disturbances and risk in running, and wherein, cascading failure is one of catastrophe failure threatening bulk power grid safe and stable operation.A lot of large area blackouts that Abroad in Recent Years occurs are all relevant with cascading failure.So-called cascading failure refers to a certain element fault in electric power system, and then causes other elements to be stopped transport, and this chain reaction rapid spread, finally causes large-scale power outage.Large-scale blackout often cause in electric power system heavy service and break down when faulty line excise the transfer causing trend in electrical network, power flow transfer is by the load of other transmission lines in further heavy system and cause the generation of cascading trip situation.How effectively to prevent the cascading trip event that in electrical network, circuit causes because power flow transfer causes overload, be prevent the important ring occurred of having a power failure on a large scale.

Summary of the invention

For the deficiency that prior art exists, the invention provides a kind of suppressing method of power grid cascading fault.

Technical scheme of the present invention:

A suppressing method for power grid cascading fault, is characterized in that: comprise the steps:

Step 1: determine the topological structure of electricity grid network and statistics network node total number b and branch of a network sum s, the branch of a network determining the branch of a network cut because of fault, power flow transfer occurs;

Step 2: set up the node importance contribute matrix of branch of a network electric current and calculate this matrix;

Generator in this electricity grid network, load are all represented as node Injection Current, then the node importance contribute matrix of branch of a network electric current is:

In formula, R (γ)=YA ^tb ^-1for b*s rank matrix, Y is the branch admittance matrix of network, and A is the node incidence matrices of network, and B is the node admittance matrix of network; w _jfor the weight coefficient of node, j=1,2 ..., s, c _jfor the cost coefficient of node; Matrix I _m(λ) what represent is the relation importance degree of branch current and node Injection Current, i.e. the significance level that affects branch current of node Injection Current, matrix I _m(λ) the first row element λ in _1j, row k element λ _kj, b row element λ _bjrepresent respectively be node 1 to node s Injection Current respectively on the significance level of Article 1 branch current impact, on the significance level of kth article branch current impact, the significance level on b article of branch current impact;

Step 3: the node provided according to the result of calculation of the node importance contribute matrix of branch of a network electric current in step 2 is on the significance level of branch current impact that power flow transfer occurs, determine to need to carry out the node of machine of cutting and the node of cutting load, and the node number of cutting machine is equal with the node number of cutting load;

Suppose that the branch road that power flow transfer occurs is kth bar branch road, then by matrix I _m(λ) row k element λ in _kjafter sorting according to numerical values recited: choose a corresponding node f from small to large _i(i=1 ..., n), i.e. load bus, as load bus to be cut; Meanwhile, the node g with load bus same number to be cut is chosen from big to small _i(i=1 ..., n), i.e. generator node, as machine node to be cut; The value n of described n need determine in conjunction with working experience according to on-site actual situations;

Step 4: determine initial unit shut algorithm amount;

Should be equal with cutting load amount in electricity grid network middle filling machine amount, and meet following formula:

$\left\{\begin{array}{c}{\mathrm{\ΔI}}_k=\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\ΔP}}_i\left({\mathrm{\γ}}_{{\mathrm{kg}}_i}\right|(1-j\tan {\mathrm{\θ}}_{g_i})/{\stackrel{\·}{U}}_{g_i}^{*}|-{\mathrm{\γ}}_{{\mathrm{kl}}_i}|(1-j\tan {\mathrm{\θ}}_{l_i})/{\stackrel{\·}{U}}_{l_i}^{*}\left|\right)\\ {\mathrm{\ΔP}}_i={\mathrm{\ΔP}}_{l_i}={\mathrm{\ΔP}}_{g_i}(i=1,...,n)\\ {\mathrm{\ΔP}}_i\≤{P_{l_i}}^0(i=1,...,n)\\ {\mathrm{\ΔP}}_i\≤{P_{g_i}}^0(i=1,...,n)\end{array}\right.---\left(1\right)$

In formula, Δ I _kfor the effective value of current change quantity on kth bar branch road, I _kfor the current effective value of kth bar branch road; I _ksetfor the minimum current value corresponding to the protection startup of kth bar branch road, S (S < 1) ensures that the safety factor of cascading trip does not occur branch road; for row k l in R (γ) matrix _irow, represent electric current and the load bus l of kth bar branch road _iinjection Current between numerical relation; for row k g in R (γ) matrix _irow, represent electric current and the generator node g of kth bar branch road _iinjection Current between numerical relation; with represent load bus l respectively _iload power factor angle and generator node g _igenerator power factor angle; with represent load bus l _iwith generator node g _ithe conjugation of voltage vector; be respectively load bus l _i, generator node g _iresection; All told for load bus l _iall told, for generator node g _iall told.

Step 5: the initial unit shut algorithm amount that step 4 obtains is optimized, obtains the optimal value of unit shut algorithm amount;

$\underset{{\mathrm{\&Delta;P}}_i}{\min }{J}_T=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}L(I(t+1),{\mathrm{\&Delta;P}}_i(t+1))+F\left(I(t+1)\right)---\left(2\right)$

s.t.

I(t+1)＝f(I(t),ΔP ₁(t),...,ΔP _n(t))

ΔP _i(t+1)∈[0,ΔP _imax](i＝1,...,n)

In formula, J is optimization object function, and I (t) and I (t+1) are respectively kth bar branch road at the current value of t and the current value in t+1 moment after carrying out unit shut algorithm; L (I, Δ P _i) represent and the predeterminated target of unit shut algorithm amount optimal control be generally taken as normal form: wherein, R and Q is the symmetrical weighting matrix of positive definite; F represents the functional relation of controlled quentity controlled variable and input variable; Δ P _i(t+1) what represent is t+1 moment unit shut algorithm amount; Δ P _imaxrepresent maximum permission resection; F (I) is terminal penalty function, and represent the state of moment t and the deviation of set point, general type is F (I)=I ^tpI, wherein, P is terminal weighting matrix.

Step 6: the electricity grid network of unit shut algorithm amount optimal value to current time obtained according to step 5 cuts machine and cutting load;

Step 7: detect the current value of the branch of a network of subsequent time generation power flow transfer and record this current value, judging whether its electric current enters in safe range, no, then go to step 8, be, then go to step 10;

Step 8: judge that unit shut algorithm amount optimizes the number of times whether number of times reaches setting in advance, no, then go to step 2, be, then go to step 9;

Step 9: the current value exporting the branch of a network that power flow transfer occurs;

Step 10: terminate.

Beneficial effect: method of the present invention is: calculate the maximum node of overload branches affect, determine that unit shut algorithm weight range is repeatedly optimized it, according to the unit shut algorithm amount after optimizing each time, machine of cutting and cutting load operation are implemented to electricity grid network, and whether the current value detecting subsequent time overload branch road is in safe range, continuous circulation carries out the optimization of unit shut algorithm amount and implements cutting load operation in this approach, until the current value of overload branch road reaches safe range or unit shut algorithm amount optimization number of times exceedes maximum times stopping.Method of the present invention can optimize unit shut algorithm amount when branch road transships, and reduces loss, effectively suppresses the generation of cascading failure simultaneously.

Accompanying drawing explanation

Fig. 1 is the IEEE30 node electricity grid network structural representation of one embodiment of the present invention;

Fig. 2 is the suppressing method flow chart of the power grid cascading fault of one embodiment of the present invention;

Fig. 3 is the curent change figure of overload branch road when not carrying out unit shut algorithm in the IEEE30 node electricity grid network of one embodiment of the present invention;

Fig. 4 transships the curent change figure of branch road after the IEEE30 node electricity grid network middle filling machine cutting load of one embodiment of the present invention.

Embodiment

Below in conjunction with accompanying drawing, one embodiment of the present invention are described in further detail.

The method suppressing power grid cascading fault is implemented to the IEEE30 node electricity grid network shown in Fig. 1, as shown in Figure 2, comprises the steps:

Step 1: count this electricity grid network and comprise 30 nodes and 35 branch roads, in the network branch road l from the topological structure of the IEEE30 electricity grid network shown in Fig. 1 ₆cut, make branch road l ₁₁power flow transfer occurs causes protection to start; From the l of branch road shown in Fig. 3 ₁₁curent change curve in can find out, at branch road l ₆after cut, branch road l ₁₁electric current raise rapidly, in the 30th cycle, there is peak in electric current, and in 62 cycles, minimum appears in electric current, the overload branch road l of period ₁₁current average be 1583.7A, seriously exceed branch road l ₁₁the permission current value 1112.23A of safe operation;

Step 2: set up the node importance contribute matrix of branch of a network electric current and calculate this matrix;

Generator in this electricity grid network, load are all represented as node Injection Current, then the node importance contribute matrix of branch of a network electric current is:

In formula, R (γ)=YA ^tb ^-1for 35*30 rank matrix, Y is the branch admittance matrix of network, and A is the node incidence matrices of network, and B is the node admittance matrix of network; w _jfor the weight coefficient of node, j=1,2 ..., s, c _jfor the cost coefficient of node; Matrix I _m(λ) what represent is the relation importance degree of branch current and node Injection Current, i.e. the significance level that affects branch current of node Injection Current, matrix I _m(λ) the first row element λ in _{1, j}, the 11st row element λ _{11, j}, the 35th row element λ _{35, j}what represent respectively is that the Injection Current of node 1 to node s is respectively to Article 1 branch road (l ₁) electric current impact significance level, to Sub_clause 11 branch road (l ₁₁) electric current impact significance level, to the 35th article of branch road (l ₃₅) electric current impact significance level;

By the node importance contribute matrix I of computing network branch current _m(λ) the Sub_clause 11 branch road (l to there is power flow transfer can be selected ₁₁) the large and node that cost is relatively little of electric current impact.The I of table 1 for calculating _m(λ) the importance degree coefficient lambda of part of nodes in the 11st row in _{11, i}.

In table 1 electricity grid network, part of nodes is to Sub_clause 11 branch road (l ₁₁) importance degree coefficient

λ 11,1	0.0583+0.0037i	λ 11,7	-0.4371-0.0032i	λ 11,13	0.0474-0.0021i
						λ 11,2	0.8931-0.0017i	λ 11,8	0.3744+0.0073i	λ 11,14	-0.0083-0.0001i

λ 11,3	-0.0351-0.0008i	λ 11,10	-0.0756-0.0024i	λ 11,15	-0.0175-0.0003i
						λ 11,4	-0.3142-0.0016i	λ 11,11	0.0512-0.0032i	λ 11,16	-0.0051-0.0001i
λ 11,5	0.0971+0.0006i	λ 11,12	-0.2640-0.0017i	λ 11,17	-0.0316-0.0016i

Step 3: determine to need to carry out the node of machine of cutting and the node of cutting load;

I can be found out from the result of calculation of step 2 _m(λ) real part is much larger than imaginary part, and the Injection Current of generator is just, so the Injection Current of load is then negative, therefore by λ _{11, i}real part by after ascending sequence, ascendingly choose 3 elements, i.e. λ _11,7, λ _11,4, λ _11,12corresponding node 7,4,12 is as load bus to be cut; Next descendingly again 3 elements, i.e. λ are chosen _11,2, λ _11,8, λ _11,5corresponding node 2,8,5 is as machine node to be cut.Namely these the 3 pairs of nodes selected are to branch road l ₁₁the relatively large and node that cost is relatively little of electric current impact, unit shut algorithm operation will be carried out to these 3 pairs of nodes.

Step 4: determine unit shut algorithm weight range;

Should be equal with cutting load amount in electricity grid network middle filling machine amount, and meet following formula:

$\left\{\begin{array}{c}{\mathrm{\&Delta;I}}_{11}=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{\mathrm{\&Delta;P}}_i\left({\mathrm{\&gamma;}}_{11,g_i}\right|(1-j\tan {\mathrm{\&theta;}}_{g_i})/{\stackrel{.}{U}}_{g_i}^{*}|-{\mathrm{\&gamma;}}_{11,l_i}|(1-j\tan {{\mathrm{\&theta;}}_l}_i)/{\stackrel{.}{U}}_{l_i}^{*}\left|\right)\\ {\mathrm{\&Delta;P}}_i={{\mathrm{\&Delta;P}}_l}_i={{\mathrm{\&Delta;P}}_g}_i(i=1,...3)\\ {\mathrm{\&Delta;P}}_i\&le;{P_{l_i}}^0(i=1,...,3)\\ {\mathrm{\&Delta;P}}_i\&le;{P_{g_i}}^0(i=1,...,3)\end{array}\right.---\left(1\right)$

In formula, safety factor S=0.85 is got, I in present embodiment _{11, set}=1308.5A, Δ I ₁₁for Sub_clause 11 branch road (l ₁₁) effective value of upper current change quantity, I ₁₁for Sub_clause 11 branch road (l ₁₁) current effective value; I _{11, set}for Sub_clause 11 branch road (l ₁₁) protection start corresponding to minimum current value, S (S < 1) be that the safety factor of cascading trip does not occur guarantee branch road; for the 11st row l in R (γ) matrix _irow, represent Sub_clause 11 branch road (l ₁₁) electric current and load bus l _iinjection Current between numerical relation; for row k g in R (γ) matrix _irow, represent Sub_clause 11 branch road (l ₁₁) electric current and generator node g _iinjection Current between numerical relation; with represent load bus l respectively _iload power factor angle and generator node g _igenerator power factor angle; with represent load bus l _iwith generator node g _ithe conjugation of voltage vector; , be respectively load bus l _i, generator node g _iresection; All told for load bus l _iall told, for generator node g _iall told.

Step 5: the unit shut algorithm weight range obtained from step 4 is optimized, and obtains the optimal value of unit shut algorithm amount.As shown in table 2, the unit shut algorithm amount after first time optimization is Δ P ₁=18.53MW, Δ P ₂=7.32MW, Δ P ₃=7.11MW; The optimization object function shown in formula (2) is adopted to be optimized the unit shut algorithm weight range that step 4 obtains;

$\underset{{\mathrm{\&Delta;P}}_i}{\min }{J}_T=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}L(I(t+1),{\mathrm{\&Delta;P}}_i(t+1))+F\left(I(t+1)\right)---\left(2\right)$

s.t.

I(t+1)＝f(I(t),ΔP ₁(t),...,ΔP _n(t))

ΔP _i(t+1)∈[0,ΔP _imax](i＝1,...,n)

In formula, J is optimization object function, and I (t) and I (t+1) are respectively kth bar branch road at the current value of t and the current value in t+1 moment after carrying out unit shut algorithm; L (I, Δ P _i) represent and the predeterminated target of unit shut algorithm amount optimal control be generally taken as normal form: wherein, R and Q is the symmetrical weighting matrix of positive definite; F represents the functional relation of controlled quentity controlled variable and input variable; Δ P _i(t+1) what represent is t+1 moment unit shut algorithm amount; Δ P _imaxrepresent maximum permission resection; F (I) is terminal penalty function, and represent the state of moment t and the deviation of set point, general type is F (I)=I ^tpI, wherein, P is terminal weighting matrix.

Step 6: the unit shut algorithm amount optimal value obtained according to step 5 is implemented to cut machine and cutting load to the electricity grid network of current time;

Step 7: the Sub_clause 11 branch road l detecting subsequent time generation power flow transfer ₁₁current value, judge whether its electric current enters in safe range, Sub_clause 11 branch road l ₁₁the permissible value of safe operation electric current be S*I _{11, set}=1112.23A, no, then go to step 8, be, then go to step 10; As shown in table 2, after having implemented first time unit shut algorithm, branch road l ₁₁current value be 1295.5A, branch road l ₁₁electric current do not enter safe range;

Step 8: judge that unit shut algorithm amount optimizes the number of times whether number of times reaches setting in advance, no, then go to step 2, be, then go to step 9, optimizing number of times to the unit shut algorithm amount of IEEE30 electricity grid network setting in present embodiment is 5 times, after present embodiment has carried out 3 unit shut algorithm amounts optimization to IEEE30 electricity grid network, as can be seen from Figure 4, in the 50th cycle, carry out first time unit shut algorithm after-current value decline rapidly, carrying out second, after three unit shut algorithms, branch road 11 current value is 1083.5A, enter in the scope of safe operation electric current, second, three unit shut algorithms, as shown in table 2, the machine node of cutting is node 8, 1, cutting load node is node 4, 7, second, three times unit shut algorithm amount optimal value is respectively Δ P ₁=3.67MW, Δ P ₂=3.48MW and Δ P ₁=1.58MW, Δ P ₂=0.83MW, tributary l after second time unit shut algorithm ₁₁current value be 1158.8A.

Step 9: output overloading branch road l ₁₁current value;

Table 2 IEEE30 electricity grid network unit shut algorithm Information Statistics table

Step 10: terminate.

In present embodiment, IEEE30 electricity grid network is after experienced by three unit shut algorithm amount Optimization Steps, and each unit shut algorithm amount reduces gradually, and the electric current of final overload branch road 11 enters in safe scope, inhibits the generation of cascading failure.

Claims (2)

1. a suppressing method for power grid cascading fault, is characterized in that: comprise the steps:

Step 1: determine the topological structure of electricity grid network and statistics network node total number b and branch of a network sum s, the branch of a network determining the branch of a network cut because of fault, power flow transfer occurs;

Step 2: set up the node importance contribute matrix of branch of a network electric current and calculate this matrix;

Generator in this electricity grid network, load are all represented as node Injection Current, then the node importance contribute matrix of branch of a network electric current is:

In formula, R (γ)=YA ^tb ^-1for b*s rank matrix, Y is the branch admittance matrix of network, and A is the node incidence matrices of network, and B is the node admittance matrix of network; w _jfor the weight coefficient of node, j=1,2 ..., s, c _jfor the cost coefficient of node; Matrix I _m(λ) what represent is the relation importance degree of branch current and node Injection Current, i.e. the significance level that affects branch current of node Injection Current, matrix I _m(λ) the first row element λ in _1j, row k element λ _kj, b row element λ _bjrepresent respectively be node 1 to node s Injection Current respectively on the significance level of Article 1 branch current impact, on the significance level of kth article branch current impact, the significance level on b article of branch current impact;

Step 3: the node provided according to the result of calculation of the node importance contribute matrix of branch of a network electric current in step 2, on the significance level of branch current impact that power flow transfer occurs, is determined to need to carry out the node of machine of cutting and the node of cutting load;

Suppose that the branch road that power flow transfer occurs is kth bar branch road, then by matrix I _m(λ) row k element λ in _kjafter sorting according to numerical values recited: choose n corresponding node l from small to large _i(i=1 ..., n), i.e. load bus, as load bus to be cut; Meanwhile, the node g with load bus same number to be cut is chosen from big to small _i(i=1 ..., n), i.e. generator node, as machine node to be cut;

Step 4: determine unit shut algorithm weight range;

Should be equal with cutting load amount in electricity grid network middle filling machine amount, and meet following formula:

$\left\{\begin{array}{c}\mathrm{\&Delta;}{I}_k=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}\mathrm{\&Delta;}{P}_i\left({\mathrm{\&gamma;}}_{{\mathrm{kg}}_i}\right|(1-j\tan {\mathrm{\&theta;}}_{g_i})/{\stackrel{.}{U}}_{g_i}^{*}|-{\mathrm{\&gamma;}}_{{\mathrm{kl}}_i}|(1-j\tan {\mathrm{\&theta;}}_{l_i})/{\stackrel{.}{U}}_{l_i}^{*}\left|\right)\\ \mathrm{\&Delta;}{P}_i=\mathrm{\&Delta;}{P}_{l_i}=\mathrm{\&Delta;}{P}_{g_i}(i=1,...,n)\\ \mathrm{\&Delta;}{P}_i\&le;{P_{l_i}}^0(i=1,...,n)\\ \mathrm{\&Delta;}{P}_i\&le;{P_{g_i}}^0(i=1,...,n)\end{array}\right.---\left(1\right)$

In formula, Δ I _kfor the effective value of current change quantity on kth bar branch road, I _kfor the current effective value of kth bar branch road; I _ksetfor the minimum current value corresponding to the protection startup of kth bar branch road, S (S < 1) ensures that the safety factor of cascading trip does not occur branch road; for row k l in R (γ) matrix _irow, represent electric current and the load bus l of kth bar branch road _iinjection Current between numerical relation; for row k g in R (γ) matrix _irow, represent electric current and the generator node g of kth bar branch road _iinjection Current between numerical relation; with represent load bus l respectively _iload power factor angle and generator node g _igenerator power factor angle; with represent load bus l _iwith generator node g _ithe conjugation of voltage vector; be respectively load bus l _i, generator node g _iresection; All told for load bus l _iall told, for generator node g _iall told;

Step 5: the unit shut algorithm weight range that step 4 obtains is optimized, obtains the optimal value of unit shut algorithm amount;

Formula (2) optimization object function is adopted to be optimized the unit shut algorithm weight range that step 4 obtains;

$\underset{\mathrm{\&Delta;}{P}_i}{\min }{J}_T=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}L(I(t+1),\mathrm{\&Delta;}{P}_i(t+1))+F\left(I(t+1)\right)---\left(2\right)$

s.t.

I(t+1)＝f(I(t),ΔP ₁(t),...,ΔP _n(t))

ΔP _i(t+1)∈[0,ΔP _imax](i＝1,...,n)

In formula, J is optimization object function, and I (t) and I (t+1) are respectively kth bar branch road at the current value of t and the current value in t+1 moment after carrying out unit shut algorithm; L (I, Δ P _i) represent and the predeterminated target of unit shut algorithm amount optimal control be generally taken as normal form: L (I, Δ P _i)=I ^tqI+ Δ P _i ^tr Δ P _i, wherein, R and Q is the symmetrical weighting matrix of positive definite; F represents the functional relation of controlled quentity controlled variable and input variable; Δ P _i(t+1) what represent is t+1 moment unit shut algorithm amount; Δ P _imaxrepresent maximum permission resection; F (I) is terminal penalty function, and represent the state of moment t and the deviation of set point, general type is F (I)=I ^tpI, wherein, P is terminal weighting matrix;

Step 6: the unit shut algorithm amount optimal value obtained according to step 5 is implemented to cut machine and cutting load to the electricity grid network of current time;

Step 7: detect the current value of the branch of a network of subsequent time generation power flow transfer and record this current value, and judging whether its electric current enters in safe range, no, then go to step 8, be, then go to step 10;

Step 8: judge that unit shut algorithm amount optimizes the number of times whether number of times reaches setting in advance, no, then go to step 2, be, then go to step 9;

Step 9: the current value exporting the branch of a network that power flow transfer occurs;

Step 10: terminate.

2. the suppressing method of power grid cascading fault according to claim 1, is characterized in that: when determining unit shut algorithm weight range in described step 4, and the load of excising in network is equal with cutting machine amount need.