CN103366315B - The distribution network operating safety appraisal procedure of load restoration value is lost based on distribution network failure - Google Patents

Abstract

The present invention relates to a kind of operating safety appraisal procedure losing load restoration value based on distribution network failure, belong to dispatching automation of electric power systems and grid simulation technical field. Distribution network is carried out N-1 fault scanning by the method, dead electricity load carries out under each fault recovery and optimization and calculates the dead electricity load value that can recover; After having scanned, the cumulative dead electricity load every time recovered obtains total recovery dead electricity load value. According to total dead electricity load value, total mistake load value that may recover, total recovery load value, set up distribution network Operational Safety indicators. This wherein, fault lose load restoration optimization be consider meritorious, idle, voltage and radial operation constraint the detailed recovery and optimization model of distribution network, this MIXED INTEGER quadratic constraints programming problem adopt branch and bound method solve; The present invention based on above-mentioned optimization model, the index of the distribution network operating safety assessment of formulation, it is possible to reflect the security of current distribution network operational conditions and network structure.

Description

The distribution network operating safety appraisal procedure of load restoration value is lost based on distribution network failure

Technical field

The invention belongs to dispatching automation of electric power systems and grid simulation technical field, consider that distribution network is meritorious, idle, the load restoration strategy mathematics planning algorithm of voltage, radial constraint detailed model in particular to a kind of, and the operating safety appraisal procedure based on this kind of recovery algorithms.

Background technology

In distribution network, dead electricity load after fault often can realize telegram in reply by automatic or manual, therefore considers to recover dead electricity load situation after the system fault of controlling factor, more can characterize the security that distribution network currently runs. The method of existing distribution network safety or reliability assessment or do not consider that the impact of load restoration or its Restoration model are very simple: the research having under N-1 fault scanning using isolation after lose load as finally losing load, it does not have consider that part or all of dead electricity load can be resumed; Some research, taking transforming plant main transformer or taking feeder line entirety as research object, have ignored the concrete structure of feeder line inside.

What traditional distribution network failure lost that load restoration method adopts mostly is heuristic method, intelligent algorithm etc., these methods or optimality cannot meet or computing time too long; Some mathematical optimization method has done great simplification because Solve problems is complicated, such as, only consider active balance and radial constraint when distribution recovers, or adding heuristic rule wherein and make computational short cut, the severity of its model and calculation result cannot ensure.

Document " ConvexModelsofDistributionSystemReconfiguration " is carrying out have employed detailed model in distribution network network reconfiguration research, but this algorithm loses in load restoration at distribution network failure and uses, and introduce more variable when research is considered network radial constraint. Document " ImposingRadialityConstraintsinDistributionSystemOptimiza tionProblems " have studied the treatment process of the radial constraint of distribution network.

Summary of the invention

It is an object of the invention to the weak point in order to overcome prior art, a kind of distribution network operating safety appraisal procedure losing load restoration value based on distribution network failure is proposed, when the advantage that the present invention has is to distribution network safety assessment, consider distribution network failure and lose load restoration, more can characterize physical security, and the security of distribution network can be reflected from network structure and current operational conditions two aspects; Recovery algorithms wherein is then adopt detailed model and mathematical programming method for solving, ensures the severity understood.

The distribution network operating safety appraisal procedure losing load restoration value based on distribution network failure that the present invention proposes, it is characterised in that, the method comprises the following steps:

1) the optimization model of band MIXED INTEGER two constraints is built:

1.1) initial recovery and optimization model is built

It is obtain by solving an optimization model that distribution network failure loses load restoration value, and the model that this optimization model is set up to recover more dead electricity loads as target, builds initial optimization model as follows:

$\mathrm{Obj}.\mathrm{Min}\underset{{i\∈\mathrm{\Φ}}_{\mathrm{out}}}{\mathrm{\Σ}}(L_{p,i}^0-\underset{j\∈i}{\mathrm{\Σ}}{x}_{\mathrm{ij}}{P}_{\mathrm{ij}})---\left(1\right)$

x_ij∈{0,1},i,j∈Φ_all(2)

$\underset{j\∈i}{\mathrm{\Σ}}{x}_{\mathrm{ij}}{P}_{\mathrm{ij}}=L_{p,i}^0,i\∈{\mathrm{\Φ}}_{\mathrm{all}}-{\mathrm{\Φ}}_{\mathrm{loss}}---\left(3\right)$

$\underset{j\∈i}{\mathrm{\Σ}}{x}_{\mathrm{ij}}{.Q}_{\mathrm{ij}}=L_{p,i}^0,i\∈{\mathrm{\Φ}}_{\mathrm{all}}-{\mathrm{\Φ}}_{\mathrm{loss}}---\left(4\right)$

$\underset{i\≠j}{\mathrm{\Σ}}{x}_{\mathrm{ij}}=N_{\mathrm{node}}-N_{\mathrm{root}},i,j\∈{\mathrm{\Φ}}_{\mathrm{all}}---\left(5\right)$

$\underset{j\∈i}{\mathrm{\Σ}}{x}_{\mathrm{ij}}{P}_{\mathrm{ij}}\≥0,i\∈{\mathrm{\Φ}}_{\mathrm{loss}}---\left(6\right)$

$L_{p,i}^0-\underset{j\∈i}{\mathrm{\Σ}}{x}_{\mathrm{ij}}{P}_{\mathrm{ij}}\≥0,i\∈{\mathrm{\Φ}}_{\mathrm{loss}}---\left(7\right)$

$\frac{P_{\mathrm{ij}}^2+Q_{\mathrm{ij}}^2}{V_i^2}\≤{\stackrel{\‾}{I}}_{\mathrm{ij}}^2,i,j\∈{\mathrm{\Φ}}_{\mathrm{all}},i\≠j---\left(8\right)$

$V_i^2-V_j^2=2(R_{\mathrm{ij}}{P}_{\mathrm{ij}}+Q_{\mathrm{ij}}{X}_{\mathrm{ij}})-(R_{\mathrm{ij}}^2+X_{\mathrm{ij}}^2)\frac{P_{\mathrm{ij}}^2+Q_{\mathrm{ij}}^2}{V_i^2},i,j\∈{\mathrm{\Φ}}_{\mathrm{all},i\≠}j---\left(9\right)$

$\underset{\‾}{V_i}\≤V_i\≤{\stackrel{\‾}{V}}_i,i\∈{\mathrm{\Φ}}_{\mathrm{all}}---\left(10\right)$

Wherein, N_nodeFor the total number of node; N_branchFor the total number of branch road; N_rootFor feeder line root node number; Φ_lossFor the load node set of dead electricity after fault isolation; Φ_allFor node set all in network; J ∈ i is the node showing to have the relation of connection with node i;WithFor the load at node i place before fault is gained merit and idle numerical value; x_ijFor the connection state of branch road in model, 0 and 1 represents disconnection respectively and closes; P_ijAnd Q_ijFor the meritorious and reactive power flow numerical value on branch road ij;For the safety current limit value of branch road ij; R_ijAnd X_ijFor resistance and the reactance value of branch road ij; V_i,It is respectively the voltage magnitude of node i, upper voltage limit, lower voltage limit.

In this initial optimization model, formula (1) is objective function, it is desired to lose load few as much as possible; Formula (2)-(10) are constraint; Formula (2) is 01 variable of reflection branch road opening and closing situation; (3)-(4) are power balance constraint; (3-5) radial constraint is jointly constituted; (6)-(7) are dead electricity node recovery power constraint; (8) it is the thermally-stabilised constraint of circuit; (9)-(10) are branch road both end voltage relation and node voltage constraint.

1.2) the partially restrained form in initial optimization model is processed:

A) (8) will be retrained to be improved to:

$P_{\mathrm{ij}}^2+Q_{\mathrm{ij}}^2\≤x_{\mathrm{ij}}.{\stackrel{\‾}{I}}_{\mathrm{ij}}^2.V_0^2,i,j\∈{\mathrm{\Φ}}_{\mathrm{all}}---(8\text{'})$

WhereinFor constant,For the branch current upper limit, V₀For node predeterminated voltage, it it is all constant;

After formula (8) is improved, the form that formula (3)-(4) are multiplied with variable different in (6)-(7): x_ij·P_ijWith x_ij·Q_ijAlso P it is denoted respectively as_ijAnd Q_ij:

$\underset{j\∈i}{\mathrm{\Σ}}{P}_{\mathrm{ij}}=L_{p,i}^0,i\∈{\mathrm{\Φ}}_{\mathrm{all}}-{\mathrm{\Φ}}_{\mathrm{loss}}---(3\text{'})$

$\underset{j\∈i}{\mathrm{\Σ}}{Q}_{\mathrm{ij}}=L_{q,i}^0,i\∈{\mathrm{\Φ}}_{\mathrm{all}}-{\mathrm{\Φ}}_{\mathrm{loss}}---(4\text{'})$

$\underset{j\∈i}{\mathrm{\Σ}}{P}_{\mathrm{ij}}\≥0,i\∈{\mathrm{\Φ}}_{\mathrm{loss}}---(6\text{'})$

$L_{p,i}^0-\underset{j\∈i}{\mathrm{\Σ}}{P}_{\mathrm{ij}}\≥0,i\∈{\mathrm{\Φ}}_{\mathrm{loss}}---(7\text{'})$

Described optimization model is made to become to solve;

B) retrain (9) to change into:

U_i-U_j=2 (P_ijR_ij+Q_ijX_ij)(9’)

i,j∈Φ_all,U_i≥0,U_j≥0

Wherein (quadratic term in formula (9) is much smaller than once item, it is possible to ignore);

C) constraint (9 ') of branch road both end voltage relation is processed further:

Introduce variable M_ij:

M_ij=(1-x_ij)·M₀(11)

M wherein₀It is a given enough big positive number, and then (9 ') is improved to:

U_i-U_j≤M_ij+2(P_ijR_ij+Q_ijX_ij)

U_i-U_j≥-M_ij+2(P_ijR_ij+Q_ijX_ij); (9 ")

i,j∈Φ_all,U_i≥0,U_j≥0

D.) constraint (3) (5) are processed, to solve the problem that zero injection node produces:

Set up an enough little load ε to all zero injection nodes, namely for the node not having dead electricity, if its initial load is 0, then establishAnd for dead electricity node, constraint (6) changes into:

$\underset{j\∈i}{\mathrm{\Σ}}{P}_{\mathrm{ij}}\≥\mathrm{\ϵ},i\∈{\mathrm{\Φ}}_{\mathrm{loss}}---(6\text{'})$

By step 1.2) improve the optimization model that rear original initial optimization model becomes band linear goal, two constraints of band MIXED INTEGER;This optimization model, taking (1) as optimization aim, is constrained to (2) (3 ') (4 ') (5) (6 ') (7 ') (8 ') (9 ' ') (11);

2) lose load restoration value based on distribution network failure to be assessed by the operating safety of distribution network:

Concrete steps are as follows:

2.1) initialize of network: distribution network is reduced to circuit to be opened/closed, and the node with load.

2.2) i=1 is made;

2.3) arranging i-th circuit is fault circuit;

2.4) fault isolation:

Fault circuit i is isolated, disconnects by all circuits being directly connected with fault circuit i, so that the node of fault circuit and association thereof becomes isolated island; The circuit disconnect these and fault circuit, and the node of association, remove from distribution network network; After fault isolation, distribution network network is divided into two parts, and a part is alive part, and another part is the dead electricity part comprising one or more dead electricity region; Dead electricity region is further divided into the dead electricity region that may recover and the dead electricity region that can not recover;

Obtain for the network structure calculated, total dead electricity load value and the dead electricity load value that may recover by the circuit in alive part and the dead electricity region that may recover of dead electricity part and node;

2.5) the optimization model of band MIXED INTEGER two constraints is solved

The input of the optimization model solving two constraints of band MIXED INTEGER is the circuit obtained in 2.1-2.4 and node network structure, node load value, line current limit and node voltage limit, the dead electricity part that this network structure only comprises alive part and may recover. Solve the optimization model of band MIXED INTEGER two constraints with branch and bound method, obtain electric network fault and lose load restoration value; Namely the actual recovery dead electricity load value in current i-th line fault situation is obtained;

2.6) " N-1 " fault scanning judges:

If i=N, N are the total number of circuit, then fault scanning completes, carries out step 2.7); If i < N, fault scanning does not complete, and namely also has other circuit to need to calculate, makes i=i+1, return to step 2.3);

2.7) distribution network safety assessment result is calculated:

It is defined as follows distribution network Operational Safety indicators as distribution network safety assessment result:

$\mathrm{Index}1=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{L}_{\mathrm{restorable},i}/\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{L}_{\mathrm{outage},i}$

$\mathrm{Index}2=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{L}_{\mathrm{restorable},i}/\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{L}_{\mathrm{outage},i}$

Wherein,For each total dead electricity load sum in fault scanning;For the dead electricity load sum that in fault scanning, each time may be recovered;The actual dead electricity load sum that can recover of for calculating by recovery algorithms each time;

Jointly characterize the operating safety of distribution network by Index1 and Index2, it is final safety assessment result.

The inventive method has the following advantages:

(1) load of dead electricity considers the recovery of dead electricity load when evaluating distribution network operating safety, so more can characterize the security that distribution network currently runs, because can partly or entirely restore electricity via recovery policy after fault.

(2) recovery policy optimization algorithm considers the detailed model of distribution network and solves by mathematic programming methods, and model is strict, can obtain optimum solution.

(3) two evaluation indexes characterize distribution network network structure and operational conditions respectively for the impact of security, embody the security of distribution network when the first two aspect for ground comprehensively.

Accompanying drawing explanation

Fig. 1 is distribution network safety assessment schema.

Fig. 2 be 69 node system N-1 scan after, total dead electricity load with may recover load value comparison diagram.

Fig. 3 is after the actual 113 node 4 feeder line distribution network system N-1 in somewhere scan, total dead electricity load with may recover load value comparison diagram.

Fig. 4 is 69 node systems when load is different, and N-1 may recover load and the actual situation figure recovering load after scanning.

Embodiment

The present invention propose a kind of based on distribution network failure lose load restoration value distribution network operating safety appraisal procedure by reference to the accompanying drawings and embodiment be described in detail as follows:

The distribution network operating safety appraisal procedure losing load restoration value based on distribution network failure that the present invention proposes, it is characterized in that, the recovery of dead electricity load is considered when distribution network is carried out safety assessment, and losing the strict mathematical optimization model that load restoration have employed distribution network detailed model, evaluation index can reflect distribution network network structure and the security of operational conditions two aspects.

The method comprises the following steps:

1) the optimization model of band MIXED INTEGER two constraints is built

1.1) initial recovery and optimization model is built

It is obtain by solving an optimization model that distribution network failure loses load restoration value, and the model that this optimization model is set up to recover more dead electricity loads as target, builds initial optimization model as follows:

$\mathrm{Obj}.\mathrm{Min}\underset{{i\&Element;\mathrm{\&Phi;}}_{\mathrm{out}}}{\mathrm{\&Sigma;}}(L_{p,i}^0-\underset{j\&Element;i}{\mathrm{\&Sigma;}}{x}_{\mathrm{ij}}{P}_{\mathrm{ij}})---\left(1\right)$

x_ij∈{0,1},i,j∈Φ_all(2)

$\underset{j\&Element;i}{\mathrm{\&Sigma;}}{x}_{\mathrm{ij}}{P}_{\mathrm{ij}}=L_{p,i}^0,i\&Element;{\mathrm{\&Phi;}}_{\mathrm{all}}-{\mathrm{\&Phi;}}_{\mathrm{loss}}---\left(3\right)$

$\underset{j\&Element;i}{\mathrm{\&Sigma;}}{x}_{\mathrm{ij}}{.Q}_{\mathrm{ij}}=L_{p,i}^0,i\&Element;{\mathrm{\&Phi;}}_{\mathrm{all}}-{\mathrm{\&Phi;}}_{\mathrm{loss}}---\left(4\right)$

$\underset{i\&NotEqual;j}{\mathrm{\&Sigma;}}{x}_{\mathrm{ij}}=N_{\mathrm{node}}-N_{\mathrm{root}},i,j\&Element;{\mathrm{\&Phi;}}_{\mathrm{all}}---\left(5\right)$

$\underset{j\&Element;i}{\mathrm{\&Sigma;}}{x}_{\mathrm{ij}}{P}_{\mathrm{ij}}\&GreaterEqual;0,i\&Element;{\mathrm{\&Phi;}}_{\mathrm{loss}}---\left(6\right)$

$L_{p,i}^0-\underset{j\&Element;i}{\mathrm{\&Sigma;}}{x}_{\mathrm{ij}}{P}_{\mathrm{ij}}\&GreaterEqual;0,i\&Element;{\mathrm{\&Phi;}}_{\mathrm{loss}}---\left(7\right)$

$\frac{P_{\mathrm{ij}}^2+Q_{\mathrm{ij}}^2}{V_i^2}\&le;{\stackrel{\&OverBar;}{I}}_{\mathrm{ij}}^2,i,j\&Element;{\mathrm{\&Phi;}}_{\mathrm{all}},i\&NotEqual;j---\left(8\right)$

$V_i^2-V_j^2=2(R_{\mathrm{ij}}{P}_{\mathrm{ij}}+Q_{\mathrm{ij}}{X}_{\mathrm{ij}})-(R_{\mathrm{ij}}^2+X_{\mathrm{ij}}^2)\frac{P_{\mathrm{ij}}^2+Q_{\mathrm{ij}}^2}{V_i^2},i,j\&Element;{\mathrm{\&Phi;}}_{\mathrm{all},i\&NotEqual;}j---\left(9\right)$

$\underset{\&OverBar;}{V_i}\&le;V_i\&le;{\stackrel{\&OverBar;}{V}}_i,i\&Element;{\mathrm{\&Phi;}}_{\mathrm{all}}---\left(10\right)$

Wherein, N_nodeFor the total number of node; N_branchFor the total number of branch road; N_rootFor feeder line root node number; Φ_lossFor the load node set of dead electricity after fault isolation; Φ_allFor node set all in network; J ∈ i is the node showing to have the relation of connection with node i;WithFor the load at node i place before fault is gained merit and idle numerical value; x_ijFor the connection state (0 and 1 represents disconnection respectively and closes) of branch road in model; P_ijAnd Q_ijFor the meritorious and reactive power flow numerical value on branch road ij;For the safety current limit value of branch road ij; R_ijAnd X_ijFor resistance and the reactance value of branch road ij; V_i, It is respectively the voltage magnitude of node i, upper voltage limit, lower voltage limit.

In this initial optimization model, formula (1) is objective function, it is desired to lose load few (the mistake load namely recovered is many as much as possible) as much as possible; Formula (2)-(10) are constraint; Formula (2) is 01 variable of reflection branch road opening and closing situation; (3)-(4) are power balance constraint; (3-5) radial constraint is jointly constituted; (6)-(7) are dead electricity node recovery power constraint; (8) it is the thermally-stabilised constraint of circuit; (9)-(10) are branch road both end voltage relation and node voltage constraint.

(owing to this initial optimization model considers the concrete circuit of distribution network, node, power and voltage, so this optimizes algorithm is the mistake load restoration algorithm for the concrete model of distribution network. )

1.2) the partially restrained form in initial optimization model is processed:

A) (8) will be retrained to be improved to:

$P_{\mathrm{ij}}^2+Q_{\mathrm{ij}}^2\&le;x_{\mathrm{ij}}.{\stackrel{\&OverBar;}{I}}_{\mathrm{ij}}^2.V_0^2,i,j\&Element;{\mathrm{\&Phi;}}_{\mathrm{all}}---(8\text{'})$

WhereinFor constant,For the branch current upper limit, V0 is node predeterminated voltage, is all constant;

(above-mentioned constraint (3)-(4) and the form having two different variablees to be multiplied in constraint (6)-(8), be equivalent to optimize negative the determining of extra large gloomy matrix of constraint in model and cannot solve, so needing formula (8) form to be improved), after formula (8) is improved, the form that formula (3)-(4) are multiplied with variable different in (6)-(7): x_ij·P_ijWith x_ij·Q_ijAlso P it is denoted respectively as_ijAnd Q_ij:

$\underset{j\&Element;i}{\mathrm{\&Sigma;}}{P}_{\mathrm{ij}}=L_{p,i}^0,i\&Element;{\mathrm{\&Phi;}}_{\mathrm{all}}-{\mathrm{\&Phi;}}_{\mathrm{loss}}---(3\text{'})$

$\underset{j\&Element;i}{\mathrm{\&Sigma;}}{Q}_{\mathrm{ij}}=L_{q,i}^0,i\&Element;{\mathrm{\&Phi;}}_{\mathrm{all}}-{\mathrm{\&Phi;}}_{\mathrm{loss}}---(4\text{'})$

$\underset{j\&Element;i}{\mathrm{\&Sigma;}}{P}_{\mathrm{ij}}\&GreaterEqual;0,i\&Element;{\mathrm{\&Phi;}}_{\mathrm{loss}}---(6\text{'})$

$L_{p,i}^0-\underset{j\&Element;i}{\mathrm{\&Sigma;}}{P}_{\mathrm{ij}}\&GreaterEqual;0,i\&Element;{\mathrm{\&Phi;}}_{\mathrm{loss}}---(7\text{'})$

Described optimization model is made to become to solve;

B) retrain (9) to change into:

U_i-U_j=2 (P_ijR_ij+Q_ijX_ij)(9’)

i,j∈Φ_all,U_i≥0,U_j≥0

WhereinUi≥0、U_jQuadratic term in >=0(formula (9) is much smaller than once item, it is possible to ignore);

C) constraint (9 ') of branch road both end voltage relation is processed further:

Introduce variable M_ij:

M_ij=(1-x_ij)·M₀(11)

M wherein₀It is a given enough big positive number being greater than 10 times of average voltages, and then (9 ') is improved to:

U_i-U_j≤M_ij+2(P_ijR_ij+Q_ijX_ij)

U_i-U_j≥-M_ij+2(P_ijR_ij+Q_ijX_ij)(9”)

i,j∈Φ_all,U_i≥0,U_j≥0

(according to constraint (9 '), when branch road i-j disconnects, the active reactive value of branch road is 0, so the equal (U of the voltage magnitude at branch road two ends_i-U_j=0), this is irrational.Variable M is introduced for addressing this problem_ij; When branch road i-j connects, x_ij=1, M_ij=0, and constraint (9 ' ') consistent with constraint (9 '); When branch road i-j disconnects, x_ij=0, M_ij=M₀, make constraint (9 ' ') become invalid constraint and not work);

D.) constraint (3) (5) are processed, to solve the problem that zero injection node produces:

Set up the enough little load ε of less than load average more than 1000 times to all zero injection nodes, namely for the node not having dead electricity, if its initial load is 0, then establishAnd for dead electricity node, constraint (6) changes into:

$\underset{j\&Element;i}{\mathrm{\&Sigma;}}{P}_{\mathrm{ij}}\&GreaterEqual;\mathrm{\&epsiv;},i\&Element;{\mathrm{\&Phi;}}_{\mathrm{loss}}---(6\text{'})$

(constraint (3)-(4) and (5), respectively with the number relation of power balance and node and circuit, form the constraint of the radial operation of distribution network jointly. But when there is the zero injection node that node load is 0 in network, constraint (3)-(4) ensure that the function of Connectivity will lose efficacy, so constraint (3)-(5) are processed; )

The optimization model of band linear goal, two constraints of band MIXED INTEGER is become by optimization model original initial after above-mentioned improvement; Optimization model after improvement, taking (1) as optimization aim, is constrained to (2) (3 ') (4 ') (5) (6 ') (7 ') (8 ') (9 ' ') (11);

2) lose load restoration value based on distribution network failure to be assessed by the operating safety of distribution network: as shown in Figure 1, concrete steps are as follows for this steps flow chart:

2.1) initialize of network: distribution network is reduced to circuit to be opened/closed, and the node with load.

2.2) i=1 is made;

2.3) arranging i-th circuit is fault circuit;

2.4) fault isolation:

Fault circuit i is isolated, disconnects by all circuits being directly connected with fault circuit i, so that the node of fault circuit and association thereof becomes isolated island; The circuit disconnect these and fault circuit, and the node of association, remove from distribution network network; After fault isolation, distribution network network is divided into two parts, and a part is alive part, and another part is the dead electricity part comprising one or more dead electricity region; And dead electricity region be further divided into the dead electricity region that may recover and can not recover dead electricity region (according to each piece of dead electricity region whether likely by connection to alive part);

Obtain for the network structure calculated, total dead electricity load value and the dead electricity load value that may recover by the circuit in alive part and the dead electricity region that may recover of dead electricity part and node;

2.5) solve mistake load restoration and optimize model

Solving the input losing load restoration optimization model is the circuit obtained in 2.1-2.4 and node network structure, node load value, line current limit and node voltage limit, the dead electricity part that this network structure only comprises alive part and may recover. Solve the optimization model of band MIXED INTEGER two constraints with branch and bound method, obtain electric network fault and lose load restoration value; Namely the actual recovery dead electricity load value in current i-th line fault situation is obtained;

2.6) " N-1 " fault scanning judges:

If i=N, N are the total number of circuit, then fault scanning completes, carries out step 2.7); If i < N, fault scanning does not complete, and namely also has other circuit to need to calculate, makes i=i+1, return to step 2.3);

2.7) distribution network safety assessment result is calculated:

It is defined as follows distribution network Operational Safety indicators as distribution network safety assessment result:

$\mathrm{Index}1=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{L}_{\mathrm{restorable},i}/\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{L}_{\mathrm{outage},i}$

$\mathrm{Index}2=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{L}_{\mathrm{restorable},i}/\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{L}_{\mathrm{outage},i}$

Wherein,For each total dead electricity load sum in fault scanning;For the dead electricity load sum that in fault scanning, each time may be recovered;The actual dead electricity load sum that can recover of for calculating by recovery algorithms each time.

It is described as follows as distribution network safety assessment result above-mentioned:

Index1 with total dead electricity recovered load value divided by total dead electricity load value, gained result is between 0 to 1, it is more big that the dead electricity load that the more big explanation of value may recover accounts for total dead electricity load proportion, it is possible to the dead electricity load of recovery is only determined by network structure, thus illustrates that network structure is more safe; If being worth more little, otherwise. So Index1 characterizes the security of current distribution network network structure.

Index2 calculates the dead electricity load value summation of actual energy recovery with recovery and optimization, divided by total dead electricity load value, gained result is between 0 to 1, it is more big that the more big dead electricity load illustrating that reality is recovered of value accounts for total dead electricity load proportion, the actual dead electricity load recovered is determined by current system operational conditions, such as voltage level, water load equality, thus illustrate that current operational conditions is more safe; If being worth more little, otherwise. So Index2 characterizes on existing distribution network network structure basis, the security of operational conditions.

Jointly characterize the operating safety of distribution network by Index1 and Index2, it is final assessment result.

Embodiment

For inventive method implementation process is described, adopt PG&E69 node system and the somewhere 113 node 4 actual distribution network system of feeder line as embodiment.

The system that this two structures is different, carry out the calculating of Index1 respectively: after N-1 scans, dead electricity load total under each fault and the load contrast that may recover, as shown in Figures 2 and 3, wherein X-coordinate is fault sequence number, and solid line is total dead electricity load, and dotted line is the dead electricity load that possible recover, the dotted line of Fig. 3 compares the dotted line of Fig. 2, more presses close to solid line in general trend. Illustrate in Fig. 3 system, major part dead electricity load is all the dead electricity load possible being resumed, this is because Fig. 3 has multiple radial feeder line, and there is the junctor between feeder line so that what the fault isolation of nearly feeder line root caused big area is lost load and can be turned for being recovered by other feeder line; And 69 node systems corresponding to Fig. 2 are independent radial feeder lines, the big area that the fault isolation of nearly radial feeder line root causes is lost load and can not be turned for recovering by other feeder line. The security of the network structure of visible Fig. 3 system is higher than the internet security of Fig. 2 system. Network structure evaluation index value Index1=0.694 in Fig. 2, and the network structure evaluation index Index1=0.950 in Fig. 3, embody the reasonableness of network structure safety assessment.

For the system PG&E69 node system that same network structure is determined, operating safety is had impact by the current operational conditions of system. First the operational conditions that system is different is set: arrange line current limit more greatly with equivalence underloading; It is less of equivalence heavy duty that line current limit is set. As shown in Figure 4, the dead electricity load value that may recover after this system N-1 fault scanning of point-like represented by dotted arrows; The dead electricity load value of actual recovery when solid line represents underloading; The dead electricity load value of actual recovery during the heavy duty of short-term shape represented by dotted arrows. As seen from the figure, solid line, totally higher than point-like dotted line, can recover more dead electricity load when underloading is described, security is better; Heavy duty time contrary. This operation of power networks condition security in these two kinds of situations is quantized, is obtained characterizing the Index2=0.891 of the current operational conditions security of network, and Index2=0.711 during heavy duty.

Claims (1)

1. one kind is lost the distribution network operating safety appraisal procedure of load restoration value based on distribution network failure, it is characterised in that, the method comprises the following steps:

1) the optimization model of band MIXED INTEGER two constraints is built:

1.1) initial recovery and optimization model is built as follows:

x_ij∈{0,1},i,j∈Φ_all(2)

Wherein, N_nodeFor the total number of node; N_branchFor the total number of branch road; N_rootFor feeder line root node number; Φ_lossFor the load node set of dead electricity after fault isolation; Φ_allFor node set all in network; J ∈ i is the node showing to have the relation of connection with node i;WithFor the load at node i place before fault is gained merit and idle numerical value; x_ijFor the connection state of branch road in model, 0 and 1 represents disconnection respectively and closes; P_ijAnd Q_ijFor the meritorious and reactive power flow numerical value on branch road ij;For the safety current limit value of branch road ij; R_ijAnd X_ijFor resistance and the reactance value of branch road ij; V_i,V _i,It is respectively the voltage magnitude of node i, upper voltage limit, lower voltage limit;

In this initial optimization model, formula (1) is objective function, it is desired to lose load few as much as possible; Formula (2)-(10) are constraint; Formula (2) is 01 variable of reflection branch road opening and closing situation; (3)-(4) are power balance constraint; (3-5) radial constraint is jointly constituted; (6)-(7) are dead electricity node recovery power constraint; (8) it is the thermally-stabilised constraint of circuit; (9)-(10) are branch road both end voltage relation and node voltage constraint;

1.2) the partially restrained form in initial optimization model is processed:

A) (8) will be retrained to be improved to:

WhereinFor constant,For the branch current upper limit, V₀For node predeterminated voltage, it it is all constant;

After formula (8) is improved, the form that formula (3)-(4) are multiplied with variable different in (6)-(7): x_ij·P_ijWith x_ij·Q_ijAlso P it is denoted respectively as_ijAnd Q_ij:

Described optimization model is made to become to solve;

B) retrain (9) to change into:

Wherein U_i=V_i ²,U_i≥0、

C) constraint (9 ') of branch road both end voltage relation is processed further:

Introduce variable M_ij:

M_ij=(1-x_ij)·M₀(11)

M wherein₀It is a given enough big positive number being greater than 10 times of average voltages, and then (9 ') is improved to:

D.) constraint (3) (5) are processed, to solve the problem that zero injection node produces:

Set up the enough little load ε of less than load average more than 1000 times to all zero injection nodes, namely for the node not having dead electricity, if its initial load is 0, then establishAnd for dead electricity node, constraint (6) changes into:

By step 1.2) improve the optimization model that rear original initial optimization model becomes band linear goal, two constraints of band MIXED INTEGER; This optimization model, taking (1) as optimization aim, is constrained to (2) (3 ') (4 ') (5) (6 ') (7 ') (8 ') (9 ") (11);

2) lose load restoration value based on distribution network failure to be assessed by the operating safety of distribution network:

Concrete steps are as follows:

2.1) initialize of network: distribution network is reduced to circuit to be opened/closed, and the node with load;

2.2) i=1 is made;

2.3) arranging i-th circuit is fault circuit;

2.4) fault isolation:

Being isolated by fault circuit i, after fault isolation, distribution network network is divided into two parts, and a part is alive part, and another part is the dead electricity part comprising one or more dead electricity region; Dead electricity region is further divided into the dead electricity region that may recover and the dead electricity region that can not recover;

Obtain for the network structure calculated, total dead electricity load value and the dead electricity load value that may recover by the circuit in alive part and the dead electricity region that may recover of dead electricity part and node;

2.5) the optimization model losing two constraints of load restoration band MIXED INTEGER is solved

Using step 2.1)-2.4) in the circuit that obtains and node network structure, node load value, line current limit and node voltage limit as the input of the optimization model of band MIXED INTEGER two constraints, the dead electricity part that this network structure only comprises alive part and may recover; Solve the optimization model of band MIXED INTEGER two constraints with branch and bound method, obtain electric network fault and lose load restoration value; Namely the actual recovery dead electricity load value in current i-th line fault situation is obtained;

2.6) " N-1 " fault scanning judges:

If i=N, N are the total number of circuit, then fault scanning completes, carries out step 2.7); If i < N, fault scanning does not complete, and namely also has other circuit to need to calculate, makes i=i+1, return to step 2.3);

2.7) distribution network safety assessment result is calculated:

It is defined as follows distribution network Operational Safety indicators as distribution network safety assessment result:

Wherein,For each total dead electricity load sum in fault scanning;For the dead electricity load sum that in fault scanning, each time may be recovered;The actual dead electricity load sum that can recover of for calculating by recovery algorithms each time; Jointly characterize the operating safety of distribution network by Index1 and Index2, it is final safety assessment result.