Beruflich Dokumente
Kultur Dokumente
Received 17 June 2003; received in revised form 19 September 2003; accepted 6 October 2003
Abstract
Radial distribution reconfiguration problem formulation is rather difficult from the computer simulation point of view. Complexity of
distribution network increases with increase in numbers of laterals and sublaterals. This needs a generalized approach, by which any radial
distribution system can be solved. This paper describes the formulation of the reconfiguration problem using network topology approach.
Algorithm developed can be used in general to radial system with any number of bifurcations. A modified iterative load flow method is also
discussed to assist the reconfiguration problem. Simple and efficient technique is described to detect the loops formed during reconfiguration
process. Formulation of single loop optimization problem is implemented to obtain network reconfiguration under normal operation.
© 2003 Elsevier B.V. All rights reserved.
∗ Corresponding author. The network topology proves its usefulness in the analysis
E-mail address: vngohokar@rediffmail.com (V.N. Gohokar). of power system due to its ability to represent the system
0378-7796/$ – see front matter © 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.epsr.2003.10.005
306 V.N. Gohokar et al. / Electric Power Systems Research 69 (2004) 305–310
into matrix form, by which the solution can be obtained very Using (1), it is easy to calculate the twigs current as given
easily. The formation of the appropriate network matrix is in (2).
an integral part of a digital computer program for solution
of power system problem. ib = kILoad (2)
In a radial distribution system, it is rather difficult to ob-
This formulation assumes that the links are not added in
tain the system condition. If the laterals and sublaterals are
the system (i.e. original system configuration). Formulation
many in number, load flow calculations and identifying the
of the path incidence matrix helps to identify the path from
system loops involve lot of computational work. These draw-
a particular node to reference node. Total system losses can
backs can be overcome by using the network topology to a
be obtained using (3)
particular radial system. This also improves the flexibility in
numbering the nodes and branches. Again the application of TLoss = iTb R̃ib (3)
KCL and KVL is quite easy, as compared to conventional
approaches. This approach does not require obtaining the R̃ is the b × b diagonal matrix of resistances of the feeders.
nodes beyond particular branches for load flow studies. It The identification of the loop formed due to addition of
also helps in obtaining the loop formation due to closing of tie switch, i.e. link can be done by formulating the loop
switches in the system to reduce losses. incidence matrix for the branches and links. Eq. (4) helps in
To reconfigure the system, it is necessary to obtain the tree identifying the loops formed in the complete system.
and co-tree of a particular radial distribution system. This
tree and co-tree formulation indeed identifies the connectiv- −(Al kT )T
C= (4)
ity of the nodes and branches and formation of the loop. In ul
the analysis of electrical network, formulation of a suitable
where C is the loop incidence matrix; Al the Bus incidence
model is the first step and model must describe the character-
matrix for the links; ul the unity matrix for links; k the path
istics of individual network components. A network matrix
incidence matrix for the branches.
equation provides a convenient model for digital simulation.
Consider primary distribution network shown in Fig. 1
KCL can be applied easily by developing the bus inci-
to illustrate (4). CB1–CB5 are normally closed sectionaliz-
dence matrix (A) for the twigs in the system. KCL at every
ing switches and CB6–CB7 are normally open Tie switches.
node is given by (1)
Two main feeders are outgoing from substation (SS). Vari-
AT ib = ILoad (1) ous nodes, branches and links are numbered as mentioned
in figure. When the tie switches are closed, two closed
where A is the b × n Bus incidence matrix considering the loops are formed which are identified by loop incidence
source node as the reference node; ib the b × 1 branch or matrix C.
twigs current matrix; ILoad the n × 1 load current matrix; b Graph of primary network is shown in Fig. 2. Matrix A,
the number of twigs or branches, n (number of nodes) − 1. Al , k and transpose of matrix C (CT ) are as given below.
4. Obtain the voltages of each node other than slack node given by (11). The loads are represented by constant current
using (6). representation model, which in turn do not alter the loop
5. Correct the PLoss and QLoss to new voltage and repeat currents. As opening of branch changes the load flow profile
steps (3) and (4). of the system, it is necessary to obtain load flow at each
6. Stop the calculations if voltage difference between two iterative stage.
consecutive iterations is within specified tolerance.