CN103346565B - Method for identifying weak nodes of power grid based on vector digraph - Google Patents

Abstract

The invention discloses a method for identifying weak nodes of a power grid based on a vector digraph in the technical field of voltage stability control of power systems. The method comprises the following steps of: determining first-level characteristic attributes of load nodes and second-level characteristic attributes respectively corresponding to the first-level characteristic attributes; acquiring values of the characteristic attributes of the load nodes; extracting main characteristic attributes of the load nodes; calculating weights of the main characteristic attributes of the load nodes; generating the vector digraphs of the load nodes according to the weight of the main characteristic attribute of each load node; calculating vector quantized values of the load nodes according to the vector digraphs of the load nodes; and determining the weak nodes of the power grid according to the vector quantized values of the load nodes. According to the method, the comprehensive attributes of the load nodes are considered, the mutual relation among the load nodes is established through the vector digraphs, and the weak nodes of the power grid are further determined, so that the efficient support is provided for the voltage stability control.

Description

Based on the Weak Node in Power Grid discrimination method of vectorial directed graph

Technical field

The invention belongs to power system voltage stabilization control technology field, particularly relate to a kind of Weak Node in Power Grid discrimination method based on vectorial directed graph.

Background technology

Along with the continuous expansion of Power System Interconnection scale, electricity needs sharp increase, transregional day by day remarkable with feature that is remote conveying electric system.In recent years, the a lot of voltage collapse accident of domestic and international generation, cause great economic loss, find by analyzing these power system accidents: electrical network generation Voltage Instability is often from some or that certain several voltage stability is relatively weak node, and then feed through to other nodes, finally cause the collapse of whole system voltage.In electric power system, different load buses has different influence degrees to voltage stability.By selecting suitable node to carry out reactive power compensation, the voltage stability of electric power system can be improved to a great extent.Here said " suitable node " is in fact exactly load bus relatively weak in electrical network.By determining load bus weak in electrical network, realizing the reactive power compensation to these nodes, and then guaranteeing the stability of power system voltage.Therefore, load bus weak in accurate recognition electrical network, for Network Voltage Stability provides guidance to become current problem demanding prompt solution.

In the prior art, the method for weak node in some identification electrical networks is had.Such as, application number is 201110091237.1, name is called the Chinese invention patent (publication No.: CN102157938A) of " power system voltage stabilization weak node ONLINE RECOGNITION method ", provide a kind of weak node ONLINE RECOGNITION method, utilize the node measurement information of discontinuity surface when network impedance data and list, realize the identification of voltage stabilization weak node.For another example, name is called the article (" electric power network technique " of " the multiple criteria integrated voltage stability index of identification Weak Node in Power Grid ", 26-31 page, Gao Peng, Shi Libao, Yao Liangzhong, Ni Yixin, Masoud Bazargan, 33rd volume the 19th phase, in November, 2009), propose a kind of discrimination method of Weak Node in Power Grid, the method according to the change in voltage index of P-V and Q-V curve calculation egress and reactive power nargin, then adopts ideal point evaluation method that two kinds of indexs are comprehensively used for identification weak node.Certainly, the method for other identification weak nodes is additionally provided in prior art.But, there is a common problem in these methods, namely only node performance is in systems in which considered according to the individual attribute of node, do not consider node performance in systems in which according to the synthesized attribute of node, more do not consider the relative performance of node from this angle of the correlation between node and node.

The present invention proposes a kind of Weak Node in Power Grid discrimination method based on vectorial directed graph, consider the synthesized attribute of node, set up the correlation between node and node by vectorial directed graph, and then determine node relative performance in systems in which, and obtain the weak node of electrical network thus.

Summary of the invention

The object of the invention is to, a kind of Weak Node in Power Grid discrimination method based on vectorial directed graph is provided, for solving existing Weak Node in Power Grid recognition technology Problems existing.

To achieve these goals, the technical scheme that the present invention proposes is that a kind of Weak Node in Power Grid discrimination method based on vectorial directed graph, is characterized in that described method comprises:

Step 1: determine the first order characteristic attribute of load bus and second level characteristic attribute corresponding to each first order characteristic attribute;

Step 2: the value obtaining the characteristic attribute of load bus, if namely the first order characteristic attribute of load bus has corresponding second level characteristic attribute, then obtains the value of second level characteristic attribute; If the first order characteristic attribute of load bus does not have corresponding second level characteristic attribute, then obtain the value of first order characteristic attribute;

Step 3: the main characteristic attribute extracting load bus;

Step 4: the weights of the main characteristic attribute of calculated load node;

Step 5: according to the weights of the main characteristic attribute of each load bus, generates load bus vector directed graph;

Step 6: according to load bus vector directed graph calculated load knot vector quantized value;

Step 7: judge Weak Node in Power Grid according to load bus vector quantization value.

The main characteristic attribute of described extraction load bus is the main characteristic attribute of union as the load bus second level characteristic attribute of setting and the value of characteristic attribute being greater than the characteristic attribute of set point.

The weights of the main characteristic attribute of described calculated load node are, the value of setting second level characteristic attribute identical in the main characteristic attribute of each load bus is added up, again divided by the minimum value in all accumulation results, as the weights of the second level characteristic attribute set in main characteristic attribute; Be 1 by the weights assignment of the characteristic attribute beyond the second level characteristic attribute set in main characteristic attribute.

Described step 5 specifically, to any two load bus g _iand g _jif, load bus g _iwith load bus g _jthere is identical main characteristic attribute, and load bus g _ithe value of identical main characteristic attribute be less than load bus g _jthe value of identical main characteristic attribute, then have one from load bus g _ipoint to load bus g _jand the directed edge of the weights of the weights main characteristic attribute that to be this identical; Wherein, i ≠ j.

Described calculated load knot vector quantized value comprises:

Sub-step 101: provide each load bus g _ivector quantization value computing formula, described computing formula is

$v_i=(1-\mathrm{\ρ})+\mathrm{\ρ}\×(\frac{n_{1,i}}{c_1}{v}_1+\frac{n_{2,i}}{c_2}{v}_2+...+\frac{n_{i-1,i}}{c_{i-1}}{v}_{i-1}+\frac{n_{i+1,i}}{c_{i+1}}{v}_{i+1}+...+\frac{n_{k,i}}{c_k}{v}_k);$

Wherein, v _iload bus g _ivector quantization value, i=1,2 ..., k;

ρ is the smoothing factor of setting, and ρ ∈ (0,1);

C _jload bus g _jpoint to the directed edge of other all load bus weights and, j=1,2 ..., k and j ≠ i;

N _j,iload bus g _jpoint to load bus g _idirected edge weights and, j=1,2 ..., k and j ≠ i;

K is the number of load bus;

Sub-step 102: by the vector quantization value computing formula simultaneous of all load buses, obtain the vector quantization value system of linear equations of load bus, solve the vector quantization value that described system of linear equations obtains each load bus.

Described according to load bus vector quantization value judge Weak Node in Power Grid be by all load buses according to its vector quantization value ascending order sequence, the preceding load bus that sorts is Weak Node in Power Grid.

The first order characteristic attribute of described load bus comprises the reactive power compensation amount of the measurement baric flow of load bus, the power of load bus and load bus;

The second level characteristic attribute that the measurement baric flow of described load bus is corresponding comprises the measurement voltage of load bus and the measured current of load bus;

The second level characteristic attribute that the power of described load bus is corresponding comprises the power-factor angle of the active power of load bus, the reactive power of load bus and load bus.

The present invention considers the synthesized attribute of load bus, sets up the correlation between load bus and load bus by vectorial directed graph, and then determines the weak node of electrical network, can provide effective support for Voltage Stability Control.

Accompanying drawing explanation

Fig. 1 is the Weak Node in Power Grid discrimination method flow chart based on vectorial directed graph;

Fig. 2 is that the load bus vector directed graph that embodiment provides generates schematic diagram.

Embodiment

Below in conjunction with accompanying drawing, preferred embodiment is elaborated.It is emphasized that following explanation is only exemplary, instead of in order to limit the scope of the invention and apply.

Embodiment

Fig. 1 is the Weak Node in Power Grid discrimination method flow chart based on vectorial directed graph, and as shown in Figure 1, the Weak Node in Power Grid discrimination method based on vectorial directed graph provided by the invention comprises:

Step 1: determine the first order characteristic attribute of load bus and second level characteristic attribute corresponding to each first order characteristic attribute.

In the present embodiment, the first order characteristic attribute of load bus is the measurement baric flow r of load bus ₁, load bus power r ₂with the reactive power compensation amount r of load bus ₃.The second level characteristic attribute that the measurement baric flow of load bus is corresponding comprises the measurement voltage r of load bus ₁₁with the measured current r of load bus ₁₂.The second level characteristic attribute that the power of load bus is corresponding comprises the active power r of load bus ₂₁, load bus reactive power r ₂₂with the power-factor angle r of load bus ₂₃.The reactive power compensation amount r of load bus ₃there is no second level characteristic attribute.

Step 2: the value obtaining the characteristic attribute of load bus, if namely the first order characteristic attribute of load bus has corresponding second level characteristic attribute, then obtains the value of second level characteristic attribute; If the first order characteristic attribute of load bus does not have corresponding second level characteristic attribute, then obtain the value of first order characteristic attribute.

In the present embodiment, actual is exactly the measurement voltage r obtaining each load bus respectively ₁₁, measured current r ₁₂, active power r ₂₁, reactive power r ₂₂, power-factor angle r ₂₃with reactive power compensation amount r ₃value.These values can obtain from the existing EMS system of electric power system or SCADA system.The present embodiment Stochastic choice 3 load bus g ₁, g ₂and g ₃, the characteristic attribute value of each load bus adopts { r ₁₁, r ₁₂, r ₂₁, r ₂₂, r ₂₃, r ₃form represents, then the value of characteristic attribute that three load buses obtain is respectively: g ₁={ 0.9,0.9,0.7,0.6,0.3,0.7}, g ₂={ 0.6,0.3,0.4,0.8,0.55,0.4}, g ₃={ 0.85,0.2,0.65,0.7,0.4,0.8}.

Step 3: the main characteristic attribute extracting load bus.

Main characteristic attribute is actually the set of some characteristic attribute, and in the present invention, defining main characteristic attribute is the union that the second level characteristic attribute of setting and the value of characteristic attribute are greater than the characteristic attribute of set point.

In the present embodiment, order measures voltage r ₁₁, active power r ₂₁with reactive power r ₂₂for the second level characteristic attribute of setting, the second level characteristic attribute namely set is as { r ₁₁, r ₂₁, r ₂₂.Getting set point is 0.5, for load bus g ₁, the characteristic attribute that the value of characteristic attribute is greater than 0.5 is { r ₁₁, r ₁₂, r ₂₁, r ₂₂, r ₃, then load bus g ₁main characteristic attribute be { r ₁₁, r ₂₁, r ₂₂∪ { r ₁₁, r ₁₂, r ₂₁, r ₂₂, r ₃}={ r ₁₁, r ₁₂, r ₂₁, r ₂₂, r ₃.Similarly, for load bus g ₂, the characteristic attribute that the value of characteristic attribute is greater than 0.5 is { r ₁₁, r ₂₂, r ₂₃, then load bus g ₂main characteristic attribute be { r ₁₁, r ₂₁, r ₂₂∪ { r ₁₁, r ₂₂, r ₂₃}={ r ₁₁, r ₂₁, r ₂₂, r ₂₃.In like manner can obtain load bus g ₃main characteristic attribute be { r ₁₁, r ₂₁, r ₂₂, r ₂₃, r ₃.

Step 4: the weights of the main characteristic attribute of calculated load node.

The weights of the main characteristic attribute of calculated load node are exactly that the value of setting second level characteristic attribute identical in the main characteristic attribute by each load bus adds up, again divided by the minimum value in all accumulation results, as the weights of the second level characteristic attribute set in main characteristic attribute; Be 1 by the weights assignment of the characteristic attribute beyond the second level characteristic attribute set in main characteristic attribute.

For the present embodiment, the second level characteristic attribute set is as { r ₁₁, r ₂₁, r ₂₂.The value of setting second level characteristic attribute identical in the main characteristic attribute of each load bus being added up, is in fact exactly by load bus g ₁, g ₂and g ₃second level characteristic attribute be r ₁₁value add up, the accumulation result obtained is 2.35.By load bus g ₁, g ₂and g ₃second level characteristic attribute be r ₂₁value add up, the accumulation result obtained is 1.75.By load bus g ₁, g ₂and g ₃second level characteristic attribute be r ₂₂value add up, the accumulation result obtained is 2.1.In above-mentioned accumulation result 2.35,1.75 and 2.1, being worth minimum is 1.75.Therefore, using 2.35/1.75=1.34,1.75/1.75=1 and 2.1/1.75=1.2 as second level characteristic attribute r ₁₁, r ₂₁and r ₂₂weights.For other main characteristic attribute, weights are set to 1, i.e. main characteristic attribute r ₂₃and r ₃weights be 1.

Step 5: according to the weights of the main characteristic attribute of each load bus, generates load bus vector directed graph.

Generate vectorial directed graph specifically, to any two load bus g _iand g _jif, g _iand g _jthere is identical main characteristic attribute, and load bus g _ithe value of identical main characteristic attribute be less than load bus g _jthe value of identical main characteristic attribute, then have one from g _ipoint to g _jand the directed edge of the weights of the weights main characteristic attribute that to be this identical; Wherein, i ≠ j.

As shown in Figure 2, in the present embodiment, load bus g ₁with load bus g ₂there is identical main characteristic attribute r ₁₁, r ₂₁and r ₂₂, load bus g ₂main characteristic attribute r ₁₁value be 0.6, be less than load bus g ₁main characteristic attribute r ₁₁value 0.9, then from g ₂point to g ₁have a directed edge, the weights of this directed edge are main characteristic attribute r ₁₁weights, namely 1.34.In like manner, load bus g ₂main characteristic attribute r ₂₁value be 0.4, be less than load bus g ₁main characteristic attribute r ₂₁value 0.7, then from g ₂point to g ₁have a directed edge, the weights of this directed edge are main characteristic attribute r ₂₁weights, namely 1.In fig. 2, from g ₂point to g ₁two directed edges adopt numerical value to add and mode show, namely in Fig. 2, from g ₂point to g ₁directed edge numerical value " 1.34+1 " represent from g ₂point to g ₁comprise two directed edges, the weights of a directed edge are the weights of 1.34, directed edge is 1.Other in Fig. 2 add identical therewith with the implication represented.Load bus g ₁main characteristic attribute r ₂₂value be 0.6, be less than load bus g ₂main characteristic attribute r ₂₂value 0.8, then from g ₁point to g ₂have a directed edge, the weights of this directed edge are main characteristic attribute r ₂₂weights, namely 1.2.

According to the method, by load bus g ₁with load bus g ₃and by load bus g ₂with load bus g ₃between directed edge draw, and weights corresponding for every bar directed edge to be marked, finally generate load bus vector directed graph.

Step 6: according to load bus vector directed graph calculated load knot vector quantized value.

Calculated load knot vector quantized value comprises:

Sub-step 101: provide each load bus g _ivector quantization value computing formula as follows:

$v_i=(1-\mathrm{\ρ})+\mathrm{\ρ}\×(\frac{n_{1,i}}{c_1}{v}_1+\frac{n_{2,i}}{c_2}{v}_2+...+\frac{n_{i-1,i}}{c_{i-1}}{v}_{i-1}+\frac{n_{i+1,i}}{c_{i+1}}{v}_{i+1}+...+\frac{n_{k,i}}{c_k}{v}_k)---\left(1\right)$

In formula (1), v _iload bus g _ivector quantization value, i=1,2 ..., k.ρ is the smoothing factor of setting, and ρ ∈ (0,1).C _jload bus g _jpoint to the directed edge of other all load bus weights and, j=1,2 ..., k and j ≠ i.N _j,iload bus g _jpoint to load bus g _idirected edge weights and, j=1,2 ..., k and j ≠ i.K is the number of load bus.、

In an embodiment, setting is level and smooth because ρ=0.5.Meanwhile, can obtain according to Fig. 2, c ₁=1.2+1.2+1=3.4, c ₂=1.34+1+1.34+1=4.68, c ₃=1.34+1+1.2=3.54.n _1,2=1.2，n _1,3=1.2+1=2.2，n _2,1=1.34+1=2.34，n _2,3=1.34+1=2.34，n _3,1=1.34+1=2.34，n _3,2=1.2。

Following 3 equations can be drawn by above-mentioned formula (1):

$v_i=(1-0.5)+0.5\×(\frac{n_{2,1}}{c_2}{v}_2+\frac{n_{3,1}}{c_3}{v}_3)=0.5+0.5\×(\frac{2.34}{4.68}{v}_2+\frac{2.34}{3.54}{v}_3)---\left(2\right)$

$v_2=(1-0.5)+0.5\×(\frac{n_{1,2}}{c_1}{v}_1+\frac{n_{3,2}}{c_3}{v}_3)=0.5+0.5\×(\frac{1.2}{3.4}{v}_1+\frac{1.2}{3.54}{v}_3)---\left(3\right)$

$v_3=(1-0.5)+0.5\×(\frac{n_{\mathrm{1,3}}}{c_1}{v}_1+\frac{n_{\mathrm{2,3}}}{c_2}{v}_2)=0.5+0.5\×(\frac{2.2}{3.4}{v}_1+\frac{2.34}{4.68}{v}_2)---\left(4\right)$

Sub-step 102: by the vector quantization value computing formula simultaneous of all load buses, obtain the vector quantization value system of linear equations of load bus, solve the vector quantization value that described system of linear equations obtains each load bus.

In the present embodiment, by formula (2)-(4) simultaneous, obtain following equation group:

$\left\{\begin{array}{c}{v}_1=0.5+0.5\×(\frac{2.34}{4.68}{v}_2+\frac{2.34}{3.54}{v}_3)\\ v_2=0.5+0.5\×(\frac{1.2}{3.4}{v}_1+\frac{1.2}{3.54}{v}_3)\\ v_3=0.5+0.5\×(\frac{2.2}{3.4}{v}_1+\frac{2.34}{4.68}{v}_2)\end{array}\right.---\left(5\right)$

Equation group (5) is ternary once linear equation group, existence and unique solution.By solving equation group (5), load bus g can be obtained _ivector quantization value v _i(i=1,2,3).The result solved is: load bus g ₁, vector quantization value v ₁=1.037, load bus g ₂vector quantization value v ₂=0.899, load bus g ₃vector quantization value v ₃=1.064

Step 7: judge Weak Node in Power Grid according to load bus vector quantization value.

By load bus g ₁, g ₂and g ₃according to its vector quantization value v ₁, v ₂and v ₃ascending order sequence, its ranking results is { v ₂, v ₁, v ₃, it can thus be appreciated that load bus v ₂node the weakest in 3 load buses, load bus v ₁the node of time weakness in 3 load buses, load bus v ₃it is node the most stable in 3 load buses.Therefore, when carrying out Voltage Stability Control, first node the weakest from electrical network is started with, and carries out reactive power compensation to the weakest node, and from electrical network, time weak node is started with and then, carries out reactive power compensation to the node of secondary weakness.Carry out successively, till the stabilization of power grids.In real work, need the load bus often more than 3 investigated, 3 nodes that the present embodiment provides are just in order to illustrate implementation procedure of the present invention.More than the situation reference said process of 3 load buses, the identification of weak node also can be realized.

The above; be only the present invention's preferably embodiment, but protection scope of the present invention is not limited thereto, is anyly familiar with those skilled in the art in the technical scope that the present invention discloses; the change that can expect easily or replacement, all should be encompassed within 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 (1)

1., based on a Weak Node in Power Grid discrimination method for vectorial directed graph, it is characterized in that described method comprises:

Step 1: determine the first order characteristic attribute of load bus and second level characteristic attribute corresponding to each first order characteristic attribute;

Step 2: the value obtaining the characteristic attribute of load bus, if namely the first order characteristic attribute of load bus has corresponding second level characteristic attribute, then obtains the value of second level characteristic attribute; If the first order characteristic attribute of load bus does not have corresponding second level characteristic attribute, then obtain the value of first order characteristic attribute;

Step 3: the main characteristic attribute extracting load bus; The main characteristic attribute of described extraction load bus is the main characteristic attribute of union as the load bus second level characteristic attribute of setting and the value of characteristic attribute being greater than the characteristic attribute of set point;

Step 4: the weights of the main characteristic attribute of calculated load node; The weights of the main characteristic attribute of described calculated load node are, the value of setting second level characteristic attribute identical in the main characteristic attribute of each load bus is added up, again divided by the minimum value in all accumulation results, as the weights of the second level characteristic attribute set in main characteristic attribute; Be 1 by the weights assignment of the characteristic attribute beyond the second level characteristic attribute set in main characteristic attribute;

Step 5: according to the weights of the main characteristic attribute of each load bus, generates load bus vector directed graph, to any two load bus g _iand g _jif, load bus g _iwith load bus g _jthere is identical main characteristic attribute, and load bus g _ithe value of identical main characteristic attribute be less than load bus g _jthe value of identical main characteristic attribute, then have one from load bus g _ipoint to load bus g _jand the directed edge of the weights of the weights main characteristic attribute that to be this identical; Wherein, i ≠ j;

Step 6: according to load bus vector directed graph calculated load knot vector quantized value; Described calculated load knot vector quantized value comprises:

Sub-step 101: provide each load bus g _ivector quantization value computing formula, described computing formula is $v_i=(1-\mathrm{\ρ})+\mathrm{\ρ}\×(\frac{n_{1,i}}{c_1}{v}_1+\frac{n_{2,i}}{c_2}{v}_2+...+\frac{n_{i-1,i}}{c_{i-1}}{v}_{i-1}+\frac{n_{i+1,i}}{c_{i+1}}{v}_{i+1}+...+\frac{n_{k,i}}{c_k}{v}_k);$

Wherein, v _iload bus g _ivector quantization value, i=1,2 ..., k;

ρ is the smoothing factor of setting, and ρ ∈ (0,1);

C _jload bus g _jpoint to the directed edge of other all load bus weights and, j=1,2 ..., k and j ≠ i;

N _j,iload bus g _jpoint to load bus g _idirected edge weights and, j=1,2 ..., k and j ≠ i;

K is the number of load bus;

Sub-step 102: by the vector quantization value computing formula simultaneous of all load buses, obtain the vector quantization value system of linear equations of load bus, solve the vector quantization value that described system of linear equations obtains each load bus;

Step 7: judge Weak Node in Power Grid according to load bus vector quantization value;

Described according to load bus vector quantization value judge Weak Node in Power Grid be by all load buses according to its vector quantization value ascending order sequence, the preceding load bus that sorts is Weak Node in Power Grid;

The first order characteristic attribute of described load bus comprises the reactive power compensation amount of the measurement baric flow of load bus, the power of load bus and load bus;

The second level characteristic attribute that the measurement baric flow of described load bus is corresponding comprises the measurement voltage of load bus and the measured current of load bus;

The second level characteristic attribute that the power of described load bus is corresponding comprises the power-factor angle of the active power of load bus, the reactive power of load bus and load bus.