CN101976840B - Network topology analysis method of power system based on quasi-square of adjacency matrix - Google Patents

Abstract

The invention provides a network topology analysis method of a power system based on the quasi-square of an adjacency matrix, which has the characteristic that the programming by using a matrix method is relatively simple, and can save computer memory and calculating time. The invention is characterized in that a step of generating connection matrix modules and a step of analyzing connection matrixes by using an inverted-sequence row scanning method are the main characteristics of the invention; in the invention, connection matrixes for analyzing the network topology can be obtained only by one matrix multiplication, which improves the speed of matrix analysis, and the calculated amount is less than that in the existing matrix method, the more the vertices, the more obvious the efficiency; nodes are performed with optimized numbering in descending order of the numbers of closed switches connected with the nodes, and buses are performed with optimized numbering in descending order of the numbers of sub-circuits connected with the buses, thereby reducing the quantity of matrix multiplication. In the invention, the value of the connection matrix is kept directly by the adjacency matrix without a special intermediate matrix, which can save half of memory space on a computer, and also reduce the assignment operation of matrix.

Description

Power system network topology analyzing method based on accurate square of adjacency matrix

Technical field

The present invention relates to a kind of network topology analytical method of electric power system, particularly a kind of power system network topology analyzing method based on matrix method.

Background technology

The power system network topological analysis is a very important module in the power automatic system, and its effect is to be converted into the required Mathematical Modeling of network analysis to the physical model of electric power system.For electric power system is analyzed, must form the circuit model of reflection bus and branch road relation through the network topology analysis.

The network topology analysis comprises bus analysis and electric island analysis two parts, and the bus analysis is the synthetic bus of a set of node that connects together through close switch, and it is that a bus that connects together through branch road assembles electric island that electric island is analyzed.Though this two parts analytic target is different, method is identical, all belongs to the method that connected graph is analyzed in the graph theory.As the network topology analytical method of electric power system connected graph analytical method, mainly contain two kinds of search method and matrix methods at present.Search method is to carry out network topology through the method for search node and adjacent node annexation to analyze, simple, the understanding easily of search method principle, but it is loaded down with trivial details to programme; Matrix method is to be expressed as adjacency matrix to connection relation between nodes, then it is carried out the method that topological analysis is carried out in matrix operation, and the matrix method programming is fairly simple, but memory demand and amount of calculation are all very big.In existing matrix method, the matrix multiple number of times is n-2 time when adopting the adjacency matrix involution to demand perfection connection matrix, and the matrix multiple number of times is log when adopting quadratic method to demand perfection connection matrix ₂(n-1), amount of calculation is all very big.The applicant is to have disclosed a kind of matrix method network topological method that only just can obtain the full-mesh matrix through twice matrix multiplication operation in 201010235958.0 patent applications at application number, and analysis speed has had large increase, but still the leeway of further raising is arranged.

Summary of the invention

For overcoming the deficiency of above-mentioned matrix method, the object of the invention is exactly to propose a kind of network topology analytical method that matrix method is programmed relative characteristic of simple but also can be saved calculator memory and operation time that not only had.

For realizing above-mentioned purpose, the present invention proposes a kind of power system network topology analyzing method based on accurate square of adjacency matrix, concrete step is:

Step 1: node switch contingency table, node branch road contingency table, informational table of nodes, the bus information table of the establishment electrical network of analyzing.

Step 2: the current electric pressure sign KV=1 that will carry out the bus analysis is set, the analysis of beginning bus.

Step 3: connect the close switch number according to each node and carry out the node optimization numbering by order from big to small.

Step 4: form the adjacency matrix of reflection node through the close switch annexation.

Step 5: call the accurate quadratic method module of adjacency matrix, generate connection matrix.

Step 6: the line scanning method is analyzed connection matrix, obtains all buses in the current electric pressure.

Step 7: current electric pressure KV=KV+1 is set.

Step 8: judge whether KV counts KVS greater than total electric pressure, if KV is greater than KVS then enter into the electric island of step 9 beginning and analyze; If KV is not more than KVS, then turn back to step 3, proceed the bus analysis of new electric pressure.

Step 9: form bus branch road contingency table according to branch road two end nodes.

Step 10: carry out bus optimization numbering by order from big to small according to each bus institute chord way.

Step 11: form the adjacency matrix of reflection bus through the branch road annexation.

Step 12: call the accurate quadratic method module of adjacency matrix, generate connection matrix.

Step 13: the line scanning method is analyzed connection matrix, obtains all electric islands.

The step that connection matrix of the present invention forms module is:

Step 1: current line i=1 is set.

Step 2: be provided with as prostatitis j=1.

Step 3: judgment matrix element a _IjWhether be 0, if a _IjBe not 0, then go to step 9.

Step 4: k=1 is set.

Step 5: judge a _IkAnd a _JkWhether all be 1,, then go to step 7 if the two is not all to be 1.

Step 6: make a _Ij=1, a _Ji=1, go to step 9.

Step 7: make k=k+1.

Step 8: judge that k whether greater than number of vertex n, if k is not more than n, then turns back to step 5.

Step 9: make j=j+1.

Step 10: whether judge j greater than n,, then turn back to step 3 if j is not more than n.

Step 11: make i=i+1.

Step 12: whether judge i greater than n, if i greater than n, finishes; Otherwise turn back to step 2.

The present invention adopts backward line scanning method to analyze connection matrix, and step is:

Step 1: Group [] the array zero clearing of record node group number, present node i=n, connected subgraph group number m=0.

Step 2: judge that whether Group [i] is 0, if be 0, then enters into the node grouping situation that step 3 begins to analyze present node i place group; If be not 0, then enter into step 8, continue to analyze the grouping situation of next node.

Step 3: group number m=m+1, row j=1.

Step 4: establish Group [i]=m.

Step 5: judge connection matrix element t _IjWhether be 1,, then make Group [j]=m if be 1.

Step 6: make j=j+1.

Step 7: whether judge j less than i, if j less than i, then turns back to step 5.

Step 8: make i=i-1.

Step 9: whether judge i less than 1, if i less than 1, finishes; Otherwise, turn back to step 2.

For the bus analysis, female wire size of the value representation node i of Group [i]; Analyze the electric island of the value representation bus i of Group [i] number for electric island.

The accurate quadratic method of adjacency matrix of the present invention square is the basis with matrix, whenever calculates an element a in the matrix multiple process _IjNew value, at once upgrade this matrix element a with this new value _IjAnd symmetry elements a _Ji, principle is following:

The matrix method formula of adjacency matrix involution is following:

T ^(k+1)＝T ^(k)·A (1)

In the formula: A is an adjacency matrix; T is a connection matrix; Subscript (k) representes that this matrix is communicated with matrix for the k level.

Repeat (1) can obtain the full-mesh matrix up to the n-1 of adjacency matrix power.Method through to connection matrix square can obtain the full-mesh matrix fast, and it is following to obtain the quadratic method formula thus:

T ^(2k)＝T ^(k)·T ^(k) (2)

1 grade in the formula (2) is communicated with matrix T ⁽¹⁾=A.

Employing formula (2) is when asking connection matrix, and the calculating of connection matrix element is following:

$t_{\mathrm{ij}}^{\left(2k\right)}=\underset{m=1}{\overset{n}{\mathrm{\Σ}}}{t}_{\mathrm{im}}^{\left(k\right)}{t}_{\mathrm{mj}}^{\left(k\right)}---\left(3\right)$

Because connection matrix is symmetrical, its element t _Mj=t _Jm, so formula (3) can be rewritten as:

$t_{\mathrm{ij}}^{\left(2k\right)}=\underset{m=1}{\overset{n}{\mathrm{\Σ}}}{t}_{\mathrm{im}}^{\left(k\right)}{t}_{\mathrm{jm}}^{\left(k\right)}---\left(4\right)$

Connection matrix reflects internodal annexation, is communicated with matrix for a m level, and whether what need care is to be communicated with between two nodes, need not know between two nodes it is what connection.If whenever obtain an element; Upgrade matrix element with it at once; Can be reflected in two nodes early in the connection matrix through the relation that plurality of nodes connects indirectly, help obtaining quickly the full-mesh matrix, also save the memory space of preserving new connection matrix.This pattern (4) is revised as:

$t_{\mathrm{ij}}=\underset{m=1}{\overset{n}{\mathrm{\Σ}}}{t}_{\mathrm{im}}{t}_{\mathrm{jm}}---\left(5\right)$

Formula (5) is no longer distinguished t _IjBe the element of which rank of connection matrix, can represent whether related getting final product of node i and node j.Also need not calculate simultaneously all connection matrix elements, only those are originally that 0 matrix element just needs to calculate.Also needn't be during calculating from n multiplication item of 1 to n calculating, as long as can confirm t _Ij=1 just can finish computational process in advance.

Connection matrix is a symmetrical matrix, when calculating connection matrix, and matrix element of every calculating; Not only upgrade this element, its symmetry elements are upgraded simultaneously according to symmetry, like this can the advancing updating symmetry elements; Accelerate to ask for the process of connection matrix, reduce the matrix multiplication number of times.The formula of asking for the connection matrix element according to symmetry is following:

$t_{\mathrm{ji}}=t_{\mathrm{ij}}=\underset{m=1}{\overset{n}{\mathrm{\Σ}}}{t}_{\mathrm{im}}{t}_{\mathrm{jm}}---\left(6\right)$

Through behind accurate square operation of connection matrix; Obtain the connection matrix of formula (7) form; Be that the last column of each connected graph in the connection matrix and the element of last row all are 1, use the line scanning method just can judge the connection situation of network from connection matrix last column.In order to narrate conveniently, below think that network only comprises a connected graph in the explanation, a plurality of connected graph situation can be regarded connection matrix as matrix in block form.

After trying to achieve the connection matrix shown in the formula (7), can adopt the line scanning method to confirm connected graph.Because the connection matrix that this method is tried to achieve is not the full-mesh matrix, number the connected relation that the pairing row of big node could reflect connected graph, so scanning method should begin to scan from big node number, promptly adopt backward line scanning method to confirm connected graph.

The formation method of adjacency matrix of the present invention is following:

The summit of node as figure, close switch is as the limit of figure when bus is analyzed.The diagonal entry assignment 1 of adjacency matrix, the element assignment 1 that has the limit to get in touch between the summit, the element assignment 0 that does not have the limit to get in touch between the summit.The summit of bus as figure, branch road is as the limit of figure when analyzing on electric island, and the diagonal entry assignment 1 of adjacency matrix has the element assignment 1 of limit contact between the summit, do not have the element assignment 0 of limit contact between the summit.

Node optimization numbering of the present invention is numbered node by the descending order of the close switch number that node connected, and bus optimization numbering is numbered bus by the descending order of way that bus connected.This with the sparse matrix during by the equation solution of coefficient matrix by node the ascending order of way of propping up of companys to carry out the node optimization method for numbering serial just in time opposite.

The invention has the beneficial effects as follows: compare with existing matrix method,

1,, have matrix method clear concept, programming characteristic of simple, but arithmetic speed is much faster than existing matrix method because the present invention has still utilized the advantage of matrix method.

2, the present invention only needs a matrix multiplication operation just can obtain being enough to the topological connection matrix of phase-split network, has improved the speed of matrix analysis.Carrying out the connected graph analysis with a network diagram that n summit arranged is example.Adopting the matrix multiple number of times of adjacency matrix under the multiplication algorithm worst case is n-2 time, and the demand perfection matrix multiple number of times of connection matrix of employing quadratic method is log ₂(n-1) inferior, adopting the matrix multiple number of times of patent applied for algorithm is 2 times.And the matrix multiple number of times when adopting the inventive method to analyze is 1 time.This shows that amount of calculation of the present invention is less than existing matrix method, the summit is many more, and efficient is obvious more.

3, the present invention adopts and to connect the descending order of close switch number by node and carry out node optimization and number, and by the descending order of way of propping up that bus connected bus is optimized numbering, has reduced the matrix multiplication operation amount.

4, the present invention directly preserves the value of connection matrix with adjacency matrix, does not need special intermediary matrix, can save half the Computer Storage space, has also reduced the matrix assignment operation.

Description of drawings

5 in the total accompanying drawing of the present invention.Wherein:

Fig. 1 is the flow chart of the inventive method.

Fig. 2 is the flow chart that connection matrix forms module in the inventive method.

Fig. 3 adopts backward line scanning method to analyze the flow chart of connection matrix in the inventive method.

Fig. 4 is the network diagram of first embodiment of the inventive method when adopting arbitrary number.

Fig. 5 is the network diagram of first embodiment of the inventive method when adopting serial number.

Embodiment

Be described further with 5 couples of the present invention of accompanying drawing below in conjunction with accompanying drawing 4.Fig. 4 is a simple electric power networks, comprises 1 power supply, 8 buses.Its adjacency matrix is following:

$A=\left[\begin{array}{cccccccc}1& 0& 0& 1& 0& 0& 0& 0\\ 0& 1& 0& 0& 0& 0& 1& 0\\ 0& 0& 1& 0& 1& 1& 0& 0\\ 1& 0& 0& 1& 0& 0& 0& 1\\ 0& 0& 1& 0& 1& 0& 1& 0\\ 0& 0& 1& 0& 0& 1& 0& 1\\ 0& 1& 0& 0& 1& 0& 1& 0\\ 0& 0& 0& 1& 0& 1& 0& 1\end{array}\right]$

Simple network shown in Figure 4 has been the topological diagram after bus is analyzed, and is following to the step that simple network shown in Figure 4 carries out the analysis of electric island according to the flow process of generation connection matrix shown in Figure 2:

Step 1: current line i=1 is set.Generate the 1st row of connection matrix.

Step 2: be provided with as prostatitis j=1.Generate the 1st row the 1st column element of connection matrix.

Step 3: judgment matrix element a _IjWhether be 0, if a _IjBe not 0, then go to step 9.

Because a ₁₁=1, go to step 9.

Step 9: make j=j+1.

J=1+1=2, the 1st row the 2nd column element of generation connection matrix.

Step 10: whether judge j greater than n,, then turn back to step 3 if j is not more than n.

N=8, j=2 is not more than n, turns back to step 3.

Step 3: judgment matrix element a _IjWhether be 0, if a _IjBe not 0, then go to step 9.

Because a ₁₂=0, execution in step 4.

Step 4: k=1 is set.

Step 5: judge a _IkAnd a _JkWhether all be 1,, then go to step 7 if the two is not all to be 1.

Because a ₂₁=0, a ₁₁And a ₂₁ Be 1 not all, go to step 7.

Step 7: make k=k+1.k＝1+1＝2。

Step 8: judge that k whether greater than number of vertex n, if k is not more than n, then turns back to step 5.

N=8, k=2 is not more than n, turns back to step 5.

Step 5: judge a _IkAnd a _JkWhether all be 1,, then go to step 7 if the two is not all to be 1.

Because a ₁₂=0, a ₁₂And a ₂₂ Be 1 not all, go to step 7.

Step 7: make k=k+1.k＝2+1＝3。

Step 8: judge that k whether greater than number of vertex n, if k is not more than n, then turns back to step 5.

N=8, k=3 is not more than n, turns back to step 5.

Step 5: judge a _IkAnd a _JkWhether all be 1,, then go to step 7 if the two is not all to be 1.

Because a ₁₃=0, a ₁₃And a ₂₃ Be 1 not all, go to step 7.

Step 7: make k=k+1.k＝3+1＝4。

Step 8: judge that k whether greater than number of vertex n, if k is not more than n, then turns back to step 5.

N=8, k=4 is not more than n, turns back to step 5.

Step 5: judge a _IkAnd a _JkWhether all be 1,, then go to step 7 if the two is not all to be 1.

Because a ₂₄=0, a ₁₄And a ₂₄ Be 1 not all, go to step 7.

Step 7: make k=k+1.k＝4+1＝5。

Step 8: judge that k whether greater than number of vertex n, if k is not more than n, then turns back to step 5.

N=8, k=5 is not more than n, turns back to step 5.

Step 5: judge a _IkAnd a _JkWhether all be 1,, then go to step 7 if the two is not all to be 1.

Because a ₂₅=0, a ₁₅And a ₂₅ Be 1 not all, go to step 7.

Step 7: make k=k+1.k＝5+1＝6。

Step 8: judge that k whether greater than number of vertex n, if k is not more than n, then turns back to step 5.

N=8, k=6 is not more than n, turns back to step 5.

Step 5: judge a _IkAnd a _JkWhether all be 1,, then go to step 7 if the two is not all to be 1.

Because a ₂₆=0, a ₁₆And a ₂₆ Be 1 not all, go to step 7.

Step 7: make k=k+1.k＝6+1＝7。

Step 8: judge that k whether greater than number of vertex n, if k is not more than n, then turns back to step 5.

N=8, k=7 is not more than n, turns back to step 5.

Step 5: judge a _IkAnd a _JkWhether all be 1,, then go to step 7 if the two is not all to be 1.

Because a ₁₇=0, a ₁₇And a ₂₇ Be 1 not all, go to step 7.

Step 7: make k=k+1.k＝7+1＝8。

Step 8: judge that k whether greater than number of vertex n, if k is not more than n, then turns back to step 5.

N=8, k=8 is not more than n, turns back to step 5.

Step 5: judge a _IkAnd a _JkWhether all be 1,, then go to step 7 if the two is not all to be 1.

Because a ₂₈=0, a ₁₈And a ₂₈ Be 1 not all, go to step 7.

Step 7: make k=k+1.k＝8+1＝9。

Step 8: judge that k whether greater than number of vertex n, if k is not more than n, then turns back to step 5.

N=8, k=9 be greater than n, execution in step 9.

Step 9: make j=j+1.

J=2+1=3, the 1st row the 3rd column element of generation connection matrix.So circulation is before j=8, and connection matrix the 1st row does not change.During j=8, generate the 1st row the 8th column element of connection matrix.

Step 10: whether judge j greater than n,, then turn back to step 3 if j is not more than n.

N=8, j=8 is not more than n, turns back to step 3.

Step 3: judgment matrix element a _IjWhether be 0, if a _IjBe not 0, then go to step 9.

Because a ₁₈=0, execution in step 4.

Step 4: k=1 is set.

Step 5: judge a _IkAnd a _JkWhether all be 1,, then go to step 7 if the two is not all to be 1.

Because a ₈₁=0, a ₁₁And a ₈₁ Be 1 not all, go to step 7.

Step 7: make k=k+1.k＝1+1＝2。

Step 8: judge that k whether greater than number of vertex n, if k is not more than n, then turns back to step 5.

N=8, k=2 is not more than n, turns back to step 5.

Step 5: judge a _IkAnd a _JkWhether all be 1,, then go to step 7 if the two is not all to be 1.

Because a ₈₂=0, a ₁₂And a ₈₂ Be 1 not all, go to step 7.

Step 7: make k=k+1.k＝2+1＝3。

Step 8: judge that k whether greater than number of vertex n, if k is not more than n, then turns back to step 5.

N=8, k=3 is not more than n, turns back to step 5.

Step 5: judge a _IkAnd a _JkWhether all be 1,, then go to step 7 if the two is not all to be 1.

Because a ₈₃=0, a ₁₃And a ₈₃ Be 1 not all, go to step 7.

Step 7: make k=k+1.k＝3+1＝4。

Step 8: judge that k whether greater than number of vertex n, if k is not more than n, then turns back to step 5.

N=8, k=4 is not more than n, turns back to step 5.

Step 5: judge a _IkAnd a _JkWhether all be 1,, then go to step 7 if the two is not all to be 1.

Because a ₁₄=1, a ₈₄=1, a ₁₄And a ₈₄All be 1, execution in step 6.

Step 6: make a _Ij=1, a _Ji=1, go to step 9.

Make a ₁₈=1, a ₈₁=1, go to step 9.So far, generate the 1st row of connection matrix, had only element a ₁₈=1 and a ₈₁=1 variation has taken place.

Step 9: make j=j+1.

j＝8+1＝9。

Step 10: whether judge j greater than n,, then turn back to step 3 if j is not more than n.

N=8, j=9 be greater than n, execution in step 11.

Step 11: make i=i+1.

I=1+1=2 generates the 2nd of connection matrix and goes.The result is: have only element a ₂₅=1 and a ₅₂=1 variation has taken place.Repeat this process up to accomplishing the work of producing connection matrix.

When i changed, corresponding connection matrix became 1 element such as table 1 by 0.

Table 1 connection matrix element variation table

It is following to generate connection matrix at last:

$\left[\begin{array}{cccccccc}1& 0& 0& 1& 1& 1& 1& 1\\ 0& 1& 1& 0& 1& 1& 1& 1\\ 0& 1& 1& 1& 1& 1& 1& 1\\ 1& 0& 1& 1& 1& 1& 1& 1\\ 1& 1& 1& 1& 1& 1& 1& 1\\ 1& 1& 1& 1& 1& 1& 1& 1\\ 1& 1& 1& 1& 1& 1& 1& 1\\ 1& 1& 1& 1& 1& 1& 1& 1\end{array}\right]$

Obviously following formula is not the full-mesh matrix, but the element of last several row all is 1, so last column can reflect the connected relation of system.

3) connection matrix is begun to carry out line scanning from last column, obtain the topological relation of network, all 8 buses are all in an electric island.

Fig. 5 is with the same 1 power supply, 8 buses of also comprising of Fig. 4.Simple network network topology analysis to shown in Figure 5 is following:

1) adjacency matrix A is:

$A=\left[\begin{array}{cccccccc}1& 1& 0& 0& 0& 0& 0& 0\\ 1& 1& 1& 0& 0& 0& 0& 0\\ 0& 1& 1& 1& 0& 0& 0& 0\\ 0& 0& 1& 1& 1& 0& 0& 0\\ 0& 0& 0& 1& 1& 1& 0& 0\\ 0& 0& 0& 0& 1& 1& 1& 0\\ 0& 0& 0& 0& 0& 1& 1& 1\\ 0& 0& 0& 0& 0& 0& 1& 1\end{array}\right]$

2) adjacency matrix A is called the accurate quadratic method module of adjacency matrix, it is following to generate connection matrix:

$T=\left[\begin{array}{cccccccc}1& 1& 1& 1& 1& 1& 1& 1\\ 1& 1& 1& 1& 1& 1& 1& 1\\ 1& 1& 1& 1& 1& 1& 1& 1\\ 1& 1& 1& 1& 1& 1& 1& 1\\ 1& 1& 1& 1& 1& 1& 1& 1\\ 1& 1& 1& 1& 1& 1& 1& 1\\ 1& 1& 1& 1& 1& 1& 1& 1\\ 1& 1& 1& 1& 1& 1& 1& 1\end{array}\right]$

Obviously following formula is the full-mesh matrix, and the connection matrix that visible different numberings obtain is different.But the element of last several row all is 1, so connection matrix last column can reflect the connected relation of system.

3) connection matrix is begun to carry out line scanning from last column, obtain the topological relation of network, all 8 buses are all in an electric island.

The present invention can adopt any programming language and programmed environment to realize, like C language, C++, FORTRAN, Delphi etc.Development environment can adopt Visual C++, Borland C++ Builder, Visual FORTRAN etc.Applied environment: a module that can be used as real-time systems such as EMS and distribution management system is used, and the power system analysis software that also can be used as an off-line uses.

An alternative embodiment of the invention is that the Hangzhou electrical network to a certain period carries out topological analysis:

This electrical network is a large-scale power grid, and scale was at that time: 187 at factory station, and 715 of bus section, busbar sections, 7329 on switch, 318 of transmission lines, 250 in transformer, wherein two winding transformer is 127, three winding transformations

123 of devices, 11 of series reactor branch roads, 232 of reactive power compensation electric capacity, 27 of reactive power compensation reactance;

Several 7097 of node, branch road comprise totally 825 of transmission line, transformer and series reactor branch roads.

The topological analysis result forms 957 of buses, 49 on electric island.Wherein 1 is the island of living, and the bus number is 704, and 48 dead islands comprise 122 buses altogether, and remaining is isolated bus.

Computing environment is the PC of the Intel Pentium of dominant frequency 1.10GHz, and bus is analyzed 0.125s consuming time, and 0.718s consuming time is analyzed on electric island, the 0.843s that is about total consuming time of topological analysis.

The present invention and adjacency matrix are relatively seen table 2 from multiplication algorithm, quadratic method and patent applied for (number of patent application 201010235958.0) result of calculation; It is thus clear that the inventive method and adjacency matrix involution, quadratic method are relatively; The matrix multiplication number of times obviously reduces, and computational speed significantly improves.Compare with patent applied for, speed also improves a lot.

Table 2 the present invention and existing method are relatively

Claims (1)

1. power system network topology analyzing method based on accurate square of adjacency matrix, concrete step is:

Step 1: node switch contingency table, node branch road contingency table, informational table of nodes, the bus information table of the establishment electrical network of analyzing;

Step 2: the current electric pressure sign KV=1 that will carry out the bus analysis is set, the analysis of beginning bus;

Step 3: connect the close switch number according to each node and carry out the node optimization numbering by order from big to small;

Step 4: form the adjacency matrix of reflection node through the close switch annexation;

Step 5: call the accurate quadratic method module of adjacency matrix, generate connection matrix;

Step 6: adopt backward line scanning method to analyze connection matrix, obtain all buses in the current electric pressure;

Step 7: current electric pressure KV=KV+1 is set;

Step 8: judge whether KV counts KVS greater than total electric pressure, if KV is greater than KVS then enter into the electric island of step 9 beginning and analyze; If KV is not more than KVS, then turn back to step 3, proceed the bus analysis of new electric pressure;

Step 9: form bus branch road contingency table according to branch road two end nodes;

Step 10: carry out bus optimization numbering by order from big to small according to each bus institute chord way;

Step 11: form the adjacency matrix of reflection bus through the branch road annexation;

Step 12: call the accurate quadratic method module of adjacency matrix, generate connection matrix;

Step 13: adopt backward line scanning method to analyze connection matrix, obtain all electric islands;

It is characterized in that:

The step of wherein said generation connection matrix module is:

Step 1: current line i=1 is set;

Step 2: be provided with as prostatitis j=1;

Step 3: judgment matrix element a _IjWhether be 0, if a _IjBe not 0, then go to step 9;

Step 4: k=1 is set;

Step 5: judge a _IkAnd a _JkWhether all be 1,, then go to step 7 if the two is not all to be 1;

Step 6: make a _Ij=1, a _Ji=1, go to step 9;

Step 7: make k=k+1;

Step 8: judge that k whether greater than number of vertex n, if k is not more than n, then turns back to step 5;

Step 9: make j=j+1;

Step 10: whether judge j greater than n,, then turn back to step 3 if j is not more than n;

Step 11: make i=i+1;

Step 12: whether judge i greater than n, if i greater than n, finishes; Otherwise turn back to step 2;

Described employing backward line scanning method is analyzed connection matrix, and step is:

Step 1: Group [] the array zero clearing of record node group number, present node i=n, connected subgraph group number m=0;

Step 2: judge that whether Group [i] is 0, if be 0, then enters into the node grouping situation that step 3 begins to analyze present node i place group; If be not 0, then enter into step 8, continue to analyze the grouping situation of next node;

Step 3: group number m=m+1, row j=1;

Step 4: establish Group [i]=m;

Step 5: judge connection matrix element t _IjWhether be 1,, then make Group [j]=m if be 1;

Step 6: make j=j+1;

Step 7: whether judge j less than i, if j less than i, then turns back to step 5;

Step 8: make i=i – 1;

Step 9: whether judge i less than 1, if i less than 1, finishes; Otherwise, turn back to step 2.

Images