CN100373733C - Tidal subgrid parallel computing method for power system - Google Patents

Abstract

The present invention relates to a network dividing parallel calculating method of the power flow of an electric power system, which comprises the following steps: the target electric power system is divided into a plurality of sub-networks; an associated matrix among the boundary points of the sub-networks is solved; primary solutions and an inverse matrix of the correction quantity of the node voltage of the boundary points of the sub-networks are solved; accurate solutions and correction quantity at a right item of the correction quantity of the node voltage of all the boundary points are solved; the right item of a linear equation of the sub-networks is corrected, and the accurate solutions of the node voltage are solved. According to the network dividing parallel calculating method of the power flow of large scale electric power systems of the present invention, the completeness of the original non-parallel software calculating processes can be kept, and the optimal numbers of the nodes of the original electric network are not changed. LU decomposition processes are not changed, and the solution processes of front and back substitution are not changed. Therefore, the work load of parallel development can be reduced, and the reliability of software is improved.

Description

Electric power system tide subnetting parallel calculating method

Technical field

The present invention relates to a kind of tidal current computing method of electric power system, more particularly, relate to a kind of subnetting parallel calculating method of electric power system tide.

Background technology

It is the most basic calculating of power system analysis that electric power system tide calculates, and its result of calculation can be used for representing the electrical network steady-state characteristic.Electric power system tide calculates except the important function of self, and it still is other analytical calculations of electric power system, as the basic calculation of transient stability calculating.

Electric power system tide calculating is the service conditions according to given electric network composition, parameter and elements such as generator, load, determines the calculating of electric power system each several part steady operation state parameter.Usually given service conditions has power, the pivot point voltage of each power supply and load point in the electric power system, the voltage and the phase angle of balance point.Running status variable to be asked mainly is the real part and the imaginary part of voltage magnitude and the phase angle or the voltage of each bus nodes of electrical network.

The electric power system tide calculation mathematic model is the Nonlinear System of Equations that derives according to known service conditions on power system network equation basis.As: the power balance equation formula that node voltage is represented by rectangular coordinate:

$\left.\begin{array}{c}\mathrm{\Δ}{P}_i=P_i-e_i\underset{k=1}{\overset{n}{\mathrm{\Σ}}}(G_{\mathrm{ik}}{e}_k-B_{\mathrm{ik}}{f}_k)-f_i\underset{k=1}{\overset{n}{\mathrm{\Σ}}}(G_{\mathrm{ik}}{f}_k+B_{\mathrm{ik}}{e}_k)=0\\ \mathrm{\Δ}{Q}_i=Q_i-f_i\underset{k=1}{\overset{n}{\mathrm{\Σ}}}(G_{\mathrm{ik}}{e}_k-B_{\mathrm{ik}}{f}_k)+e_i\underset{k=1}{\overset{n}{\mathrm{\Σ}}}(G_{\mathrm{ik}}{f}_k+B_{\mathrm{ik}}{e}_k)=0\\ (i=\mathrm{1,2},......n)\end{array}\right\}---\left(1\right)$

Wherein:

G _Ik, B _IkReal part and imaginary part for the admittance between node i in the admittance matrix and the node k;

P _i, Q _iMeritorious and reactive power for node i;

e _i, f _iReal part and imaginary part for the node voltage of node i.

Electric power system tide calculates can be summed up as on mathematics finds the solution Nonlinear System of Equations, and its Mathematical Modeling is write a Chinese character in simplified form as follows:

F(X)＝0 (2)

Wherein:

F=(f ₁, f ₂... f _n) ^TPower balance equation formula for node;

X=(x ₁, x ₂... x _n) ^TBe each node voltage to be asked.

Relevant trend Calculation Method is a lot, but basic, the most practical method for solving is the Newton-Raphson method (abbreviation Newton method) in conjunction with sparse matrix technology, and this method is representative method.

The Newton method iterative formula of the Nonlinear System of Equations of formula (2) is as follows:

F＇(X ^(t))ΔX ^(t)＝-F(X ^(t)) (3)

X ^(t+1)＝X ^(t)+ΔX ^(t) (4)

Wherein

$F\&prime;\left(X\left(t\right)\right)={\left[\begin{array}{ccc}\frac{\&PartialD;{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">f</figure-callout>}}_1}{\&PartialD;x_1}& ...& \frac{\&PartialD;{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">f</figure-callout>}}_1}{\&PartialD;x_n}\\ .& & .\\ .& & .\\ .& & .\\ \frac{\&PartialD;f_n}{\&PartialD;x_1}& ...& \frac{\&PartialD;f_n}{\&PartialD;x_n}\end{array}\right]}_{x-x^{\left(t\right)}}=\left[\begin{array}{ccc}{J}_{11}& ...& J_{1n}\\ .& & .\\ .& & .\\ .& & .\\ {\mathrm{<figure-callout\; id="1"\; label="J\; n"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">J</figure-callout>}}_n& ...& J_{\mathrm{nn}}\end{array}\right]$

Be Jacobian matrix, this matrix is sparse, and each element is the second order submatrix.

ΔX ^(t)＝(Δx ₁，Δx ₂…Δx _n) ^T

For revising vector.

Become the process of finding the solution system of linear equations repeatedly separating Nonlinear System of Equations like this, until ‖ Δ X ^(t)‖ → 0

To sum up, the electric power system tide computational methods can be summed up as the problem of finding the solution of large-scale sparse linear equation group, promptly

AX＝B?(5)

Wherein:

$A=\left[\begin{array}{ccc}{a}_{11}& \&CenterDot;\&CenterDot;\&CenterDot;& a_{1n}\\ \&CenterDot;& & \&CenterDot;\\ \&CenterDot;& & \&CenterDot;\\ \&CenterDot;& & \&CenterDot;\\ a_{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">n</figure-callout>}1}& \&CenterDot;\&CenterDot;\&CenterDot;& a_{\mathrm{nn}}\end{array}\right]$ Being coefficient matrix, is nonsingular Sparse Array, i.e. a Jacobian matrix;

X=(x ₁... x _n) ^TBe unknown number vector to be asked, i.e. voltage correction vector;

B=(b ₁... b _n) ^TBe known right-hand vector vector.

In the Newton method trend computational process, the iteration coefficient matrices A all becomes with voltage each time.

The general solution of sparse linear equation group is the triangle decomposition method, is about to the product that coefficient matrices A is decomposed into lower triangular matrix L and upper triangular matrix U.

A＝LU (6)

Its solution procedure is undertaken by following former generation and two steps of back substitution:

LY=B (former generation) (7)

UX=Y (back substitution) (8)

Inject unit and improve computational speed for reducing, also need to carry out node numbering optimization when calculating beginning.

The key that electric power system tide calculates solution is to find the solution the sparse linear equation group of formula (5), and its parallel algorithm also is that the parallel algorithm with system of linear equations is a foundation.

Expansion along with power system development and electric power system scale, interconnected between the Da Qu electric power system particularly, speed and efficient that the trend that is used for large-scale power system is calculated are had higher requirement, and the parallel trend of subnetting is calculated and is become the key that addresses this problem for this reason.Meanwhile, based on the overall process dynamic real-time simulator of the large-scale power system of subnetting parallel algorithm, the subnetting parallel calculating method that also requires corresponding trend to calculate adapts with it.

As mentioned above, the whole bag of tricks that electric power system tide calculates, final finding the solution of sparse linear equation group AX=B, former generation and the backward steps that this wherein needs to relate to the optimization of grid nodes numbering and coefficient matrix of linear equations A is carried out LU triangle decomposition and solving equation group AX=B of all being summed up as.

The subnetting parallel calculating method of electric power system tide adopts BBDF (Bordered Block Diagonal Form) method or similar with it method more at present.The shortcoming of these class methods is a plurality of links that need relate to computational process:

(1) when each subnet node numbering optimization, must be in the end the boundary node layout that links to each other with other subnet, with guarantee its be " edged piecemeal diagonal angle form " (BBDF).

(2) when each subnet carries out the LU triangle decomposition, need carry out boundary point gather and to the independent decomposition of boundary point piece.

(3) when each subnet carries out former generation and back substitution and finds the solution, also need aggregation process to boundary point.

Above-mentioned situation has been destroyed the complete procedure of Solving Linear, has influenced the modular construction of software, has reduced the reliability of software.

Summary of the invention

The subnetting parallel calculating method that the purpose of this invention is to provide a kind of electric power system, make it on the basis of existing electric power system tide software for calculation etc., when carrying out the parallelization reorganization, keep the complete of original computed in software process, wherein do not influence the node optimization numbering of original electrical network; Do not influence the complete procedure that the LU decomposition of each subnet and former generation and back substitution are found the solution, therefore can reduce the parallelization development workload, improve the reliability of software.

This method be applicable to electric power system tide in calculating the whole bag of tricks (Newton method, PQ decomposition method etc.) and the network equation in the transient stability find the solution, also can be used for finding the solution of general linear equation group.

According to the present invention, a kind of subnetting parallel calculating method of electric power system tide is provided, comprise the steps:

Step 1 is divided into a plurality of subnets by main control computer with target power system;

Step 2 is obtained the incidence matrices Yc between the boundary point of each subnet by main control computer;

Step 3 is obtained tentatively the separating and the inverse matrix of sub-net boundary point corresponding with it of correction of the node voltage of sub-net boundary point corresponding with it respectively by each handset, and sends to main control computer;

Step 4 is gathered earlier tentatively the separating and the inverse matrix of sub-net boundary point corresponding with it of correction of the node voltage of each sub-net boundary point by main control computer, obtain accurately the separating and the right-hand vector correction of correction of the node voltage of whole boundary points then, and send to each handset;

Each handset of step 5 is accurately separated and the right-hand vector correction according to the correction of the node voltage of the boundary point that receives from main control computer, revises the right-hand vector of each subnet system of linear equations, obtains accurately separating of each node voltage;

Each handset of step 6 judges respectively whether the calculating of subnet corresponding with it restrains, and the judged result of each handset is aggregated into main control computer, judge whether that by main control computer all-ones subnet all restrains, if the judged result of main control computer is a "No", then repeating step three is to step 5, till the calculating to all-ones subnet has all restrained.

Subnetting parallel calculating method according to large-scale electrical power system trend of the present invention can keep the complete of original non-concurrent software computational process, and this wherein original grid nodes optimization numbering is constant; The LU decomposable process is constant; The solution procedure of former generation and back substitution is constant.So just can reduce the parallelization development workload, improve the reliability of software.

In addition, described method be applicable to electric power system tide in calculating the whole bag of tricks (Newton method, PQ decomposition method etc.) and the network equation in the transient stability find the solution, also can be used for finding the solution of general linear equation group.

Description of drawings

Fig. 1 is the flow chart of electric power system tide parallel calculating method of the present invention;

Fig. 2 has schematically shown a kind of partitioning scheme of target power system and the boundary point of each subnet.

Embodiment

Generally, the High-Performance Computing Cluster machine in order to carry out the subnetting parallel computation of electric power system tide, can to adopt comprising a main control computer and a plurality of handsets.Link by network between main control computer and the handset, finish communication, control and synchronous between them.

According to the topology optimization method target power system is divided into a plurality of subnets by main control computer, and the distribution of computation tasks of each subnet is carried out to each handset, by the unified control of main control computer parallel computation.

Fig. 1 is the flow chart of the subnetting parallel calculating method of electric power system tide of the present invention.

As shown in Figure 1, the subnetting parallel calculating method according to electric power system tide of the present invention comprises the steps:

Step 101: target power system is divided into a plurality of subnets by main control computer

In this step 101, the method for being cut apart by branch road by main control computer is divided into a plurality of subnets with target power system.

Fig. 2 shows a kind of electric power system partitioning scheme.As shown in Figure 2, target power system is split into three subnet S1, S2 and S3, wherein a ₁ ⁽¹⁾Boundary point for subnet 1; a ₄ ⁽²⁾, a ₅ ⁽²⁾Boundary point for subnet 2; a ₇ ⁽³⁾Boundary point for subnet 3.

The coefficient matrix structure of the whole network equation is:

The sparse linear equation group of three subnets that are divided into is respectively,

Subnet 1:A ₁X ₁=B ₁(A ₁Be the coefficient matrix of subnet 1, X ₁Be the correction of the node voltage of subnet 1, down with)

${\mathrm{<figure-callout\; id="1"\; label="B"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">B</figure-callout>}}_1=\left[\begin{array}{c}{b}_1^{\left(1\right)}\\ b_2^{\left(1\right)}\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ b_{n_1}^{\left(1\right)}]\end{array}\right]$ ${\mathrm{<figure-callout\; id="1"\; label="X"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">X</figure-callout>}}_1=\left[\begin{array}{c}{x}_1^{\left(1\right)}\\ x_2^{\left(1\right)}\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ x_{n_1}^{\left(1\right)}\end{array}\right.]---\left(10\right)$

Subnet 2:A ₂X ₂=B ₂

${\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">B</figure-callout>}}_2=\left[\begin{array}{c}{b}_1^{\left(1\right)}\\ b_2^{\left(2\right)}\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ b_{n_2}^{\left(2\right)}\end{array}\right]$ ${\mathrm{<figure-callout\; id="2"\; label="X"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">X</figure-callout>}}_2=\left[\begin{array}{c}{x}_2^{\left(2\right)}\\ x_2^{\left(2\right)}\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ x_{n_2}^{\left(2\right)}\end{array}\right]---\left(11\right)$

Subnet 3:A ₃X ₃=B ₃

${\mathrm{<figure-callout\; id="3"\; label="B"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">B</figure-callout>}}_3=\left[\begin{array}{c}{b}_1^{\left(2\right)}\\ b_2^{\left(2\right)}\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ b_{n_3}^{\left(2\right)}\end{array}\right]$ ${\mathrm{<figure-callout\; id="3"\; label="X"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">X</figure-callout>}}_3=\left[\begin{array}{c}{x}_1^{\left(3\right)}\\ x_2^{\left(3\right)}\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ x_{n_3}^{\left(3\right)}\end{array}\right]---\left(12\right)$

Step 102: form between each sub-net boundary point incidence matrices in this step 10 by main control computer, ask incidence matrices Yc between boundary point by the subnet order:

Element wherein is the Jacobian matrix element that interconnection admittance and boundary point voltage form, and is equivalent to the associated element between subnet in formula (9) Jacobian matrix, and Yc is constant all the time in computational process.

Step 103: the node numbering optimization of carrying out each subnet by each handset respectively

In this step 103, the method for its node numbering optimization method during with non-parallel computation is identical, only does once in whole calculating.

Step 104: ask Jacobi's battle array, triangle decomposition respectively, obtain tentatively the separating and inverse matrix of correction of the node voltage of each sub-net boundary point by former generation and back substitution by handset, and be sent to main control computer;

In this step 104, at first try to achieve coefficient matrices A i, its method for solving is identical during with non-parallel computation;

Coefficient matrices A i is carried out triangle decomposition, make Ai=Li*Ui, its method is identical during also with non-parallel computation;

Ask for according to Ai*Xi '=Bi each sub-net boundary point node voltage correction tentatively separate Xi '.Wherein, can use sparse vector method and carry out former generation and back substitution, to reduce its amount of calculation.

In this step 104, the inverse matrix of each boundary point is the matrix that is made of the respective element in the inverse matrix of the coefficient matrix of each subnet, and its matrix exponent number is identical with the number of boundary point.In the present embodiment, the inverse matrix of the boundary point of each subnet is respectively:

Subnet 1: ${\mathrm{<figure-callout\; id="1"\; label="Z"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">Z</figure-callout>}}_1={\stackrel{\&OverBar;}{a}}_{11}^{\left(1\right)}$ (

Be A ₁ ^-1In corresponding element)

Subnet 2: ${\mathrm{<figure-callout\; id="2"\; label="Z"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">Z</figure-callout>}}_2=\left[\begin{array}{cc}{\stackrel{\&OverBar;}{a}}_{44}^{\left(2\right)}& {\stackrel{\&OverBar;}{a}}_{45}^{\left(2\right)}\\ {\stackrel{\&OverBar;}{a}}_{54}^{\left(2\right)}& {\stackrel{\&OverBar;}{a}}_{55}^{\left(2\right)}\end{array}\right]$ (

(i=4,5; J=4,5) be A ₂ ^-1In corresponding element)

Subnet 3: ${\mathrm{<figure-callout\; id="3"\; label="Z"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">Z</figure-callout>}}_3={\stackrel{\&OverBar;}{a}}_{77}^{\left(3\right)}$ (

Be A ₃ ^-1In corresponding element)

Ask boundary point inverse matrix Zi can use of the method realization of sparse vector method, to reduce its amount of calculation by former generation and back substitution.

The inverse matrix Zi that tentatively separates Xi ' and boundary point of the correction of the node voltage of above-mentioned boundary point must be sent to main control computer by each handset.

Step 105: gather tentatively the separating and inverse matrix of correction of the node voltage of each sub-net boundary point by main control computer, obtain accurately the separating and the right-hand vector correction of correction of the node voltage of whole boundary points, and be distributed to each handset

In this step 105, at first main control computer receives each sub-net boundary point inverse matrix Z ₁, Z ₂, Z ₃, and gather the total inverse matrix Zc of formation boundary point:

$Z_c=\left[\begin{array}{ccc}{Z}_1& & \\ & Z_2& \\ & & Z_3\end{array}\right]---\left(14\right)$

Receive then each sub-net boundary point node voltage correction tentatively separate X ₁', X ₂', X ₃', and gather each sub-net boundary point voltage of formation and tentatively separate global matrix Xc ':

$X_c^{\&prime;}=\left[\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="X"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">X</figure-callout>}}_1^{\&prime;}\\ {\mathrm{<figure-callout\; id="2"\; label="X"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">X</figure-callout>}}_2^{\&prime;}\\ {\mathrm{<figure-callout\; id="3"\; label="X"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">X</figure-callout>}}_3^{\&prime;}\end{array}\right]---\left(15\right)$

And then ask the correction amount X that tentatively separates of correction of the node voltage of each sub-net boundary point by formula (16) _c, wherein I is a unit matrix, the matrix exponent number equals the boundary point sum.In the present embodiment, its matrix exponent number is 4;

[I+Z _cY _c][ΔX _c]＝-Z _cY _cX′ _c (16)

Revise tentatively the separating of correction of the node voltage of each sub-net boundary point again, utilize formula (17) to ask accurately the separating of correction of the node voltage of each sub-net boundary point:

X _c＝X′ _c+ΔX _c (17)

Ask the right-hand vector correction of the correction of each sub-net boundary point voltage at last according to formula (18):

$\mathrm{\&Delta;}{B}_c=-Y_c{X}_c=\left[\begin{array}{c}\mathrm{\&Delta;}{\mathrm{<figure-callout\; id="1"\; label="B"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">B</figure-callout>}}_1\\ \mathrm{\&Delta;}{\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="C20041000015400141.png"\; state="\{\{state\}\}">B</figure-callout>}}_2\\ \mathrm{\&Delta;}{B}_3\end{array}\right]---\left(18\right)$

And be distributed to each subnet.

Step 106: revise the right-hand vector of each subnet system of linear equations by each handset respectively according to the right-hand vector correction of the correction of boundary point voltage, obtain accurately separating that this iterates.

In this step 106, at first revise the right-hand vector of each subnet system of linear equations:

Ai*Xi＝Bi+ΔBi (19)

And obtain accurately separating of this iteration; Revise each node voltage with Xi then; Judge at last whether iteration restrains, and the mark that will whether restrain is sent to main control computer.

Step 107:, judge whether that all-ones subnet all restrains by the whether convergence mark that main control computer receives and gathers that each subnet sends? if judged result is a "No", then repeat above-mentioned steps 104 to step 106, all restrain until all-ones subnet.

This method also is applicable to the subnetting parallel computation of electric power system transient stability, and the block parallel that also extends to any system of linear equations calculates.

In conjunction with the accompanying drawings the specific embodiment of the present invention is described above.It should be noted, the invention is not restricted to above-mentioned execution mode, under the premise of without departing from the spirit of the present invention, those skilled in the art can carry out multiple modifications and changes.

Claims (10)

1. the subnetting parallel calculating method of an electric power system tide comprises the steps:

Step 1 is divided into a plurality of subnets by main control computer with target power system;

Step 2 is obtained the incidence matrices Yc between the boundary point of each subnet by main control computer;

Step 3 is obtained tentatively the separating and the inverse matrix of sub-net boundary point corresponding with it of correction of the node voltage of its corresponding sub-net boundary point respectively by each handset, and sends to main control computer;

Step 4 is gathered earlier tentatively the separating and the inverse matrix of sub-net boundary point corresponding with it of correction of the node voltage of each sub-net boundary point by main control computer, obtain accurately the separating and the right-hand vector correction of correction of the node voltage of whole boundary points then, and send to each handset;

Each handset of step 5 is accurately separated and the right-hand vector correction according to the correction of the node voltage of the boundary point that receives from main control computer, revises the right-hand vector of each subnet system of linear equations, obtains accurately separating of each node voltage;

Each handset of step 6 judges respectively whether the calculating of subnet corresponding with it restrains, and the judged result of each handset is aggregated into main control computer, judge whether that by main control computer all-ones subnet all restrains, if the judged result of main control computer is a "No", then repeating step three is to step 5, till the calculating to all-ones subnet has all restrained.

2. the subnetting parallel calculating method of electric power system tide as claimed in claim 1, wherein, the method that described step 1 is cut apart by branch road by main control computer realizes.

3. the subnetting parallel calculating method of electric power system tide as claimed in claim 1 wherein, between described step 2 and step 3, also comprises by each handset and respectively subnet corresponding with it is carried out the step that node numbering optimization is handled.

4. the subnetting parallel calculating method of electric power system tide as claimed in claim 1, wherein, the element of described incidence matrices is the Jacobian matrix element that interconnection admittance between the boundary point of each subnet and boundary point voltage form.

5. the subnetting parallel calculating method of electric power system tide as claimed in claim 4, wherein, described incidence matrices forms according to the subnet order.

6. the subnetting parallel calculating method of electric power system tide as claimed in claim 1, wherein, the step of tentatively separating of correction of obtaining the node voltage of each sub-net boundary point in described step 3 comprises:

Obtain the coefficient matrix of subnet corresponding respectively with it by each handset;

Described coefficient matrix is carried out triangle decomposition;

Find the solution the system of linear equations of each subnet, draw tentatively the separating of voltage correction of each boundary point.

7. the subnetting parallel calculating method of electric power system tide as claimed in claim 6, wherein, the voltage correction of described each boundary point separate tentatively that sparse vector method carries out former generation and back substitution is tried to achieve by using.

8. the subnetting parallel calculating method of electric power system tide as claimed in claim 1, wherein, the inverse matrix of described boundary point is made of the respective element in the inverse matrix of the coefficient matrix of each subnet.

9. the subnetting parallel calculating method of electric power system tide as claimed in claim 8, wherein, the inverse matrix of described boundary point is by using sparse vector method and carry out former generation and back substitution being tried to achieve.

10. the subnetting parallel calculating method of electric power system tide as claimed in claim 1, wherein, described step 4 comprises:

By main control computer the boundary point inverse matrix of each subnet is aggregated into the total inverse matrix Zc of boundary point, each sub-net boundary point voltage is tentatively separated be aggregated into the total column vector Xc ' of each preliminary Xie of sub-net boundary point voltage correction,

Obtain the correction amount Xc that boundary point voltage correction is tentatively separated according to following formula,

[I+Z _cY _c][ΔX _c]＝-Z _cY _cX _c′

Wherein, I is a unit matrix, and exponent number equals the boundary point sum,

Utilize following formula to ask accurately the separating of correction of the node voltage of each sub-net boundary point:

X _c＝X _c′＋ΔX _c

Ask the right-hand vector correction of the correction of each sub-net boundary point voltage according to following formula:

$\mathrm{\&Delta;}{B}_c=-Y_c{X}_c=\left[\begin{array}{c}\mathrm{\&Delta;}{B}_1\\ \mathrm{\&Delta;}{B}_2\\ \mathrm{\&Delta;}{B}_3\end{array}\right].$

Images