CN103971026B - General method for calculating tide of positive power distribution networks - Google Patents

Abstract

The invention provides a general method for calculating the tide of positive power distribution networks. The general method comprises the step of establishing various distributed generation (DG) tide calculation models to obtain a PV node sensitivity reactance matrix BPV of nPV PV type DG positive power distribution networks so as to calculate the tide of the positive power distribution networks. The general method is high in universality, can be used by power distribution networks with various wiring modes and operating modes, and also can be used for calculating the tide of power distribution networks at disconnected areas; the topological analysis of the power distribution networks is not needed, various path matrixes do not need to be calculated, and a node admittance matrix can be directly figured out according to node information and branch information, so that the general method is simple and is easy in implementation; the accurate PV node sensitivity reactance matrix is adopted, so that the convergence speed of reactive power output of the PV node type DG is enhanced; on the basis of a Zbus method, the calculation of the whole convergence speed is fast, and the stability is good; the generable method can be expanded to calculation of three-phase tides.

Description

General method for load flow calculation of active power distribution network

Technical Field

The invention relates to a general method for load flow calculation of an active power distribution network.

Background

The power distribution network load flow calculation is an important content of power distribution system analysis, and is an important basis for quantitative analysis of rationality, reliability and economy of power distribution network system planning design and operation modes. With the rapid development of distributed power generation technology and the requirement of users on high power supply reliability, the traditional single-feeder radial power distribution network cannot meet the requirement. The future-oriented active power distribution network has the characteristics of distributed power supply, multi-feeder or single-feeder power supply and multi-loop network, and the traditional power distribution network push-back load flow calculation method is not universal to the active power distribution network any more. The access of the distributed power supply increases a new node type in the power distribution network, and the network operation mode of the power distribution network becomes complex and flexible due to the characteristics of multi-feeder or single-feeder power supply and looped network. Therefore, the research on the efficient and universal power flow calculation method of the active power distribution network has important significance.

Disclosure of Invention

The invention provides a general method for load flow calculation of an active power distribution network based on a Zbus method, aiming at the characteristics that the active power distribution network contains a distributed power supply, a multi-feeder or single-feeder power supply and a looped network and the inapplicability of the traditional load flow calculation method.

The technical solution of the invention is as follows:

a general method for calculating the load flow of an active power distribution network comprises the following steps,

s1, establishing various distributed power flow calculation models including a PQ node type DG, a PQ (V) type DG, a PI type DG and a PV type DG;

s2, obtaining n_PVPlatform PV type DG active power distribution network PV node sensitivity reactance matrix B_PV；

S3, the general method for the active power distribution network load flow calculation is as follows:

s31, determining the serial numbers of all nodes in the active power distribution network, inputting various information matrixes, forming a feeder information matrix, a node information matrix and a branch information matrix, and forming a grid-connected DG information matrix which comprises PV, PQ (V) and PI type DG information matrixes;

s32, calculating node admittance matrix Y of the whole network from the node information matrix and the branch information matrix, and calculating (n) from the node admittance matrix and the PV type DG information matrix according to the method of the step S2_PV×n_PV) Order PV node sensitivity reactance matrix B_PV；

S33, solving the injection current vector of each node of the power distribution network under the independent action of the root node of the feeder lineDeleting the row and the column of the root node in the total network node admittance matrix Y to obtain an admittance matrix Y', which is expressed by a formulaObtaining the voltage of each node under the independent action of the root node of the feeder line;

s34, initializing initial reactive power output (Q) of each PV node_max+Q_min)/2；

S35, setting the initial value of the node voltageCalculating the reactive power output of all PI and PQ (V) nodes；

S36, calculating the node voltage under the action of the injected current of each node except the root node of the feeder line according to the method in the step S33

S37, obtaining the node voltage according to the superposition principle:

s38, correcting PV node type DG reactive output Q_PV＝Q_PV+ Δ Q, check Q_PVIf not, converting the voltage into a PQ node, and recalculating a PV node sensitivity reactance matrix;

s39, checking an iteration convergence condition: all nodesReactive non-out-of-limit PV nodeThe reactive out-of-limit PV node has Q reactive output_maxOr Q_min(ii) a If the convergence condition is satisfied, the process proceeds to step S310; otherwise, go to step S36;

and S310, finishing calculation and outputting a result.

Further, the feeder information matrix format is { feeder node number; feeder node voltage value };

the node information matrix format is { node number; node load (not including DG); node-to-ground equivalent admittance parameters };

the branch information matrix format is { branch number; numbering a starting node; numbering end nodes; a branch impedance parameter; branch tie switch state };

the PV type DG information matrix format is { access node; active power output; a voltage target value; an upper reactive power output limit; lower limit of reactive power output };

the PQ type DG information matrix format is { access node; active power output; reactive power output }; the format of the PQ (V) -type DG information matrix is { access node; active power output };

the PI type DG information matrix format is { access node; active power output; a current target value; an upper reactive power output limit; lower limit of reactive power output }.

Further, the general calculation method of the PV node sensitivity reactance matrix of the active power distribution network facing the multi-feeder power supply and multi-loop network is as follows:

s21, providing n_PVThe platform PV type DG active power distribution network PV node sensitivity admittance matrix is Y_PVWhich is (n)_PV×n_PV) Order symmetric matrix, Y_PV,i,iIs the self-admittance of the ith PV node, Y_PV,i,j＝Y_PV,j,iMutual admittance for ith and j PV nodes;

S22、Y_PV,i,iwhen the feeder node and the ith PV node are grounded, the short-circuit current is generated when the unit voltage source is connected to the ith PV node;

S23、Y_PV,i,jwhen the ith PV node is connected to a unit voltage source and the jth PV node is grounded, negative short-circuit current flows;

s24 formulaB_PV＝Im(Z_PV) Determining a sensitivity reactance matrix B_PV。

Further, the PQ node type DG outputs constant active and reactive power, and in the load flow calculation, the calculation model is as follows when the load processing is regarded as negative:

wherein, P₀And Q₀Respectively PQ type DGActive and reactive outputs.

Furthermore, the pq (v) -type DG is treated as a negative load with voltage static characteristics, and during each iteration, the reactive power output of the DG needs to be updated, and the calculation model is as follows:

wherein,and in the k iteration, the DG grid connection point voltage amplitude is.

Further, the model of PI type DG is:

in load flow calculation, the PI type DG is converted into a PQ node by the following formula, namely, the load with voltage static characteristics which is regarded as negative:

wherein, I₀Constant current amplitude, Q, for PI type DG output_minAnd Q_maxRespectively being a DG reactive power lower limit and an upper limit; in each iteration process, checking whether the DG reactive power output is within the upper limit and the lower limit, and otherwise, taking the lower limit or the upper limit.

Further, the model of PV-type DG is:

taking a reactive initial value (Q) at the beginning of the load flow calculation_max+Q_min) And/2, in each iteration, obtaining a voltage difference value, and correcting the reactive power output through a sensitivity matrix.

Further, the specific steps of correcting the reactive power output are as follows:

s1, in the first iteration, the initial PV-type DG value is:

Q_DG＝-Q₀＝-(Q_max+Q_min)/2；

s2, calculating the voltage unbalance amount according to the following formula in the k iteration:

wherein, V₀As the DG voltage control target value,when the iteration is the kth time, the DG access point voltage amplitude value is obtained;

s3, correcting DG reactive power output:

in the iterative process, the PV node reactive power output is corrected according to the mismatch of the PV node voltage amplitude, and the node voltage equation:

the unit voltage is multiplied by two sides of the above formula at the same time, and the voltage angle difference can be ignored when the power distribution network line is short:

ΔV＝Z_PV(ΔP+jΔQ)

wherein,substituting the above equation, since Δ P is zero, the real parts on both sides are equal:

ΔV＝B_PVΔQ

the correction amount of the PV reactive power output can be obtained as follows:

wherein, B_PV＝Im(Z_PV) A PV node sensitivity reactance matrix;

therefore, the PV node reactive power output correction formula is as follows:

wherein, is Δ Q^kIs the correction value of the DG reactive power at the kth iteration,andand respectively providing the DG reactive power and the corrected reactive power in the k iteration, wherein in the process of each iteration, whether the reactive power of each PV type DG is within the upper limit and the lower limit is required to be checked, otherwise, the lower limit or the upper limit is selected, and the sensitivity reactance matrix is modified.

The invention has the beneficial effects that: the general method for calculating the load flow of the active power distribution network is based on a Zbus method, node voltages under the action of a root node and other nodes are respectively calculated by utilizing a superposition principle, a PV node is used as a PQ node to be processed, and reactive power output is corrected by utilizing a sensitivity reactance matrix. The invention has the following advantages:

firstly, a distribution network topology analysis is not needed, various road matrixes are not needed to be calculated, a node admittance matrix can be directly obtained through node information and branch information, and the method is simple and easy to implement;

secondly, a PV node type DG reactive output convergence speed is increased by adopting an accurate PV node sensitivity reactance matrix;

thirdly, based on the Zbus method, the overall convergence speed of calculation is high, and the stability is good;

the method is strong in universality, can be used for power distribution networks with various wiring modes and operation modes, and can also be used for load flow calculation of non-interconnected regional power distribution networks;

and fifthly, calculation of three-phase power flow can be expanded.

Drawings

Fig. 1 is a diagram of a modified IEEE33 active power distribution system.

Detailed Description

The following further illustrates a specific use method of the present invention in conjunction with a modified IEEE33 example of active power distribution system, which verifies the reliability, stability and versatility of the present invention for various forms of active power distribution networks.

A general method for load flow calculation of an active power distribution network specifically comprises the following steps:

the method comprises the following steps: establishing various distributed power flow calculation models

1) PQ node type DG

The PQ node type DG outputs constant active power and reactive power, and is treated as a load of 'negative' in load flow calculation. The double-fed induction wind power generation, photovoltaic power generation and other DGs are usually controlled by constant power factors. The calculation model is as follows:

wherein, P₀And Q₀The active and reactive outputs of the PQ-type DG are respectively;

2) PQ (V) -type DG

The PQ (V) type DG is treated as a negative load with voltage static characteristics, and the reactive power output of the DG needs to be updated in each iteration process. The calculation model is as follows:

wherein,and in the k iteration, the DG grid connection point voltage amplitude is. This model includes DG for synchronous generator interfaces and asynchronous generator interfaces without field regulation capability.

3) PI type DG

The PI type DG, commonly known as the DG of the current control power electronic converter interface, has a model of:

in the load flow calculation, the load is converted into a PQ node through the following formula, namely the load with the voltage static characteristic regarded as negative:

wherein, I₀Constant current amplitude, Q, for PI type DG output_minAnd Q_maxRespectively being a DG reactive power lower limit and an upper limit; in each iteration process, checking whether the DG reactive power output is within the upper limit and the lower limit, and otherwise, taking the lower limit or the upper limit.

4) PV type DG

PV-type DG is commonly a DG of voltage control power electronic converter interface, such as voltage control fuel cell, photovoltaic power generation, etc., and its model is:

taking a reactive initial value at the beginning of the load flow calculation, wherein the reactive initial value is generally (Q)_max+Q_min) And/2, in each iteration, obtaining a voltage difference value, and correcting the reactive power output through a sensitivity matrix. The method comprises the following specific steps:

a) for the first iteration, the initial value of the PV type DG is as follows:

Q_DG＝-Q₀＝-(Q_max+Q_min)/2

b) at the kth iteration, the voltage unbalance is calculated by:

wherein, V₀As the DG voltage control target value,for the kth iteration, DG access point voltage magnitude.

c) And (3) correcting DG reactive power output:

the correction amount of the PV reactive power output is as follows:

wherein, B_PV＝Im(Z_PV) Is the PV node sensitivity reactance matrix.

Therefore, the PV node reactive power output correction formula is as follows:

wherein, is Δ Q^kIs the correction value of the DG reactive power at the kth iteration,andthe DG reactive power output and the corrected reactive power output in the k iteration are respectively. In each iteration process, whether the reactive output of each PV type DG is within the upper limit and the lower limit or not needs to be checked, and if not, the lower limit or the upper limit is taken, and the sensitivity reactance matrix is modified.

Step two: finding n_PVPlatform PV type DG active power distribution network PV node sensitivity reactance matrix B_PV

The embodiment is directed to an active power distribution network which is possible to supply power to multiple feeders and comprises multiple loops of networks, and the general calculation method of the PV node sensitivity reactance matrix is as follows:

1) is provided with n_PVThe platform PV type DG active power distribution network PV node sensitivity admittance matrix is Y_PVWhich is (n)_PV×n_PV) Order symmetric matrix, Y_PV,i,iIs the self-admittance of the ith PV node, Y_PV,i,j＝Y_PV,j,iMutual admittance for ith and j PV nodes;

2)Y_PV,i,iwhen the feeder node and the ith PV node are grounded, the short-circuit current is generated when the unit voltage source is connected to the ith PV node;

3)Y_PV,i,jwhen the ith PV node is connected to a unit voltage source and the jth PV node is grounded, negative short-circuit current flows.

4) By the formulaB_PV＝Im(Z_PV) Determining a sensitivity reactance matrix B_PV。

When a certain PV type DG is converted into a PQ node, the reactance matrix B is adjusted to the sensitivity_PVOnly the DG at B needs to be deleted_PVThe corresponding rows and columns.

Step three: the general flow of the method for load flow calculation of the active power distribution network is as follows:

1) and determining the serial numbers of all nodes (including feeder root nodes) in the active power distribution network, and inputting various information matrixes. Forming a feeder information matrix, a node information matrix and a branch information matrix, wherein the format of the feeder information matrix is { feeder node number; feeder node voltage value }; the node information matrix format is { node number; node load (not including DG); node-to-ground equivalent admittance parameters, and the branch information matrix format is { branch number; numbering a starting node; numbering end nodes; a branch impedance parameter; branch tie switch state }. Forming a grid-connected DG information matrix comprising PV, PQ (V) and PI type DG information matrices, wherein the format of the PV type DG information matrix is { access node; active power output; a voltage target value; an upper reactive power output limit; lower limit of reactive power output }; the PQ type DG information matrix format is { access node; active power output; reactive power output }; the format of the PQ (V) -type DG information matrix is { access node; active power output }; the PI type DG information matrix format is { access node; active power output; a current target value; an upper reactive power output limit; lower limit of reactive power output }.

2) The node information matrix and the branch information matrix are used for calculating a node admittance matrix Y of the whole network, and the node admittance matrix and the PV type DG information matrix are used for calculating (n)_PV×n_PV) Order PV node sensitivity reactance matrix B_PV。

3) The injection current vector of each node of the power distribution network under the independent action of the root node of the feeder line is obtainedDeleting the row of the root node in the whole network node admittance matrix YAnd column, obtaining admittance matrix Y', which is formed by formulaAnd solving the voltage of each node under the independent action of the root node of the feeder line.

4) Initializing initial reactive output (Q) of each PV node_max+Q_min)/2。

5) Setting the initial value of the node voltageCalculating reactive power output of all PI and PQ nodes;

6) calculating the node voltage under the action of the injected current of each node (including PQ, PV and PI nodes) except the root node of the feeder line according to a similar method in the step 3)

7) Obtaining the node voltage by the superposition principle:

8) correcting PV node reactive power output Q_PV＝Q_PV+ Δ Q, check Q_PVIf not, otherwise convert to PQ node and recalculate PV node sensitivity reactance matrix (e.g., the ith PV node converts to PQ node, only delete B_PVRow i and column i).

9) Checking an iteration convergence condition: all nodesReactive non-out-of-limit PV nodeThe reactive out-of-limit PV node has Q reactive output_maxOr Q_min. Entering step 10) if the convergence condition is met; otherwise, the step 6) is carried out.

10) And (5) finishing the calculation and outputting a result.

Analysis by calculation example:

fig. 1 is a modified IEEE33 node distribution network with dashed lines for possible future active distribution system closures. The convergence accuracy of the system reference power, the reference voltage and the power flow is respectively 10MVA, 12.66KV and 10^-5。

TABLE 1 load flow calculation results for variable feeder and looped network numbers

Table 1 shows load flow calculation results of variable feeder number and ring network number, and it can be seen that the load flow calculation method of the present invention reduces the number of iterations and increases the algorithm stability with the increase of the number of feeders and ring networks.

The active output of the existing PV node type fuel cell is 300KW, the upper limit and the lower limit of the reactive output are 300KVAR and 0 respectively, the existing PV node type fuel cell is accessed to a power distribution network under different network operation modes by 4 schemes shown in a table 2, and load flow calculation results of different schemes are shown in a table 3.

TABLE 2 Fuel cell Access protocol

Table 3 load flow calculation results of PV node-containing power distribution network

Under the schemes 1 and 2, the reactive power output of the fuel cell does not reach the upper limit, and the voltage of the access point can be controlled at a target value; under the scheme 3, the voltage of the access point cannot reach a target value, and the reactive power output reaches an upper limit; under the scheme 4, the voltage of one fuel cell access point is controlled at a target value, the voltage of the other fuel cell access point cannot reach the target value, and the reactive power reaches the upper limit. Compared with different network operation modes of schemes 1, 2 and 3, the PV-type DG has the strongest control capability on the voltage of the access point under the open loop, and the PV-type DG has the weakest control capability on the voltage of the access point under the double-feed line full closed loop. The load flow calculation method can be used for carrying out load flow calculation on the power distribution network containing the PV-type nodes in various network operation modes, and the iteration times are not increased much.

Table 4 shows the scheme of merging DG of different node types into the power distribution network, and the load flow calculation is performed by using the method of the present invention under three network operations, and the result is shown in table 5. The voltage of the PV-type DG access point under the open loop cannot reach a target value, and the reactive power output reaches an upper limit; the voltage of a PV-type DG grid-connected node under the double-fed line full closed loop is larger than a target value, and the reactive power output of the PV-type DG grid-connected node is 0 lower limit.

Table 4DG access scheme

The above embodiments are only for illustrating the technical idea of the present invention, and the protection scope of the present invention is not limited thereby, and any modifications made on the basis of the technical scheme according to the technical idea of the present invention fall within the protection scope of the present invention.

Claims (3)

1. A general method for load flow calculation of an active power distribution network is characterized by comprising the following steps: comprises the following steps of (a) carrying out,

s1, establishing various distributed power flow calculation models including a PQ node type DG, a PQ (V) type DG, a PI type DG and a PV type DG;

the PQ node type DG outputs constant active power and reactive power, load processing is regarded as negative in load flow calculation, and a calculation model is as follows:

wherein, P₀And Q₀The active and reactive outputs of the PQ-type DG are respectively;

the PQ (V) type DG is treated as a negative load with voltage static characteristics, the reactive power output of the DG needs to be updated in each iteration process, and a calculation model is as follows:

wherein,when the iteration is the kth time, the DG grid connection point voltage amplitude value is obtained;

the model of PI type DG is:

in load flow calculation, the PI type DG is converted into a PQ node by the following formula, namely, the load with voltage static characteristics which is regarded as negative:

$Q_{\mathrm{DG}}=-\sqrt{{\left|I_0\right|}^2{\left|V_0^k\right|}^2-P_0^2}(Q_{\min }\≤Q_{\mathrm{DG}}\≤Q_{\max })$

wherein, I₀Constant current amplitude, Q, for PI type DG output_minAnd Q_maxRespectively being a DG reactive power lower limit and an upper limit; in each iteration process, checking whether the DG reactive power output is within an upper limit and a lower limit, or else, taking a lower limit or an upper limit;

the model of PV-type DG is:

taking a reactive initial value (Q) at the beginning of the load flow calculation_max+Q_min) And/2, in each iteration, obtaining a voltage difference value, and correcting the reactive power output through a sensitivity matrix, wherein the specific steps of correcting the reactive power output are as follows:

s1, in the first iteration, the initial PV-type DG value is:

Q_DG＝-Q₀＝-(Q_max+Q_min)/2；

s2, calculating the voltage unbalance amount according to the following formula in the k iteration:

wherein, V₀As the DG voltage control target value,when the iteration is the kth time, the DG access point voltage amplitude value is obtained;

s3, correcting DG reactive power output:

in the iterative process, PV node reactive power output is corrected according to the voltage amplitude mismatch amount of the PV node, and the correction amount of the PV reactive power output is as follows:

wherein, B_PV＝Im(Z_PV) A PV node sensitivity reactance matrix;

therefore, the PV node reactive power output correction formula is as follows:

wherein, is Δ Q^kIs the correction value of the DG reactive power at the kth iteration,andrespectively DG reactive power output and corrected reactive power output in the k iteration, wherein in each iteration process, whether the reactive power output of each PV type DG is within the upper and lower limits or not needs to be checked, otherwise, the lower limit or the upper limit is selected, and the sensitivity reactance matrix is modified

S2, obtaining n_PVPlatform PV type DG active power distribution network PV node sensitivity reactance matrix B_PV；

S3, the general method for the active power distribution network load flow calculation is as follows:

s31, determining the serial numbers of all nodes in the active power distribution network, inputting various information matrixes, forming a feeder information matrix, a node information matrix and a branch information matrix, and forming a grid-connected DG information matrix which comprises PV, PQ (V) and PI type DG information matrixes;

s32, calculating node admittance matrix Y of the whole network from the node information matrix and the branch information matrix, and calculating (n) from the node admittance matrix and the PV type DG information matrix according to the method of the step S2_PV×n_PV) Order PV node sensitivity reactance matrix B_PV；

S33, solving the injection current vector of each node of the power distribution network under the independent action of the root node of the feeder lineDeleting the row and the column of the root node in the total network node admittance matrix Y to obtain an admittance matrix Y', which is expressed by a formulaObtaining the voltage of each node under the independent action of the root node of the feeder line;

s34, initializing initial reactive power output (Q) of each PV node_max+Q_min)/2；

S35, setting the initial value of the node voltageCalculating reactive power output of all PI and PQ (V) nodes;

s36, calculating the node voltage under the action of the injected current of each node except the root node of the feeder line according to the method in the step S33

S37, obtaining the node voltage according to the superposition principle:

s38, correcting PV node type DG reactive output Q_PV＝Q_PV+ Δ Q, where Δ Q is the correction of PV reactive power, check Q_PVIf not, converting the voltage into a PQ node, and recalculating a PV node sensitivity reactance matrix;

s39, checking an iteration convergence condition: all nodesReactive non-out-of-limit PV nodeWherein, V₀The reactive output of the PV node is Q for the DG voltage control target value_maxOr Q_min(ii) a If the convergence condition is satisfied, the process proceeds to step S310; otherwise, go to step S36;

and S310, finishing calculation and outputting a result.

2. The general method for load flow calculation of an active power distribution network according to claim 1, characterized in that:

the feeder information matrix format is { feeder node number; feeder node voltage value };

the node information matrix format is { node number; node load, not including DG; node-to-ground equivalent admittance parameters };

the branch information matrix format is { branch number; numbering a starting node; numbering end nodes; a branch impedance parameter; branch tie switch state };

the PV type DG information matrix format is { access node; active power output; a voltage target value; an upper reactive power output limit; lower limit of reactive power output };

the PQ type DG information matrix format is { access node; active power output; reactive power output }; the format of the PQ (V) -type DG information matrix is { access node; active power output };

the PI type DG information matrix format is { access node; active power output; a current target value; an upper reactive power output limit; lower limit of reactive power output }.

3. The general method for load flow calculation of an active power distribution network according to claim 1, characterized in that: the general calculation method of the PV node sensitivity reactance matrix of the active power distribution network facing multi-feeder power supply and multi-loop network is as follows:

s21, providing n_PVThe platform PV type DG active power distribution network PV node sensitivity admittance matrix is Y_PVWhich is (n)_PV×n_PV) Order symmetric matrix, Y_PV,i,iIs the self-admittance of the ith PV node, Y_PV,i,j＝Y_PV,j,iMutual admittance for ith and j PV nodes;

S22、Y_PV,i,iwhen the feeder node and the ith PV node are grounded, the short-circuit current is generated when the unit voltage source is connected to the ith PV node;

S23、Y_PV,i,jwhen the ith PV node is connected to a unit voltage source and the jth PV node is grounded, negative short-circuit current flows;

s24 formulaB_PV＝Im(Z_PV) Determining a sensitivity reactance matrix B_PV。