| Abstract | Real power losses in distribution systems in general are quite appreciable, constituting major
portion of the overall power system losses. Power distribution systems especially in
developing countries are characterised by increased power losses, poor voltage profile,
inadequate metering, frequent failure of major equipment mainly due to non-availability of
their loading information and lack of proper monitoring and co-ordinated controls.
Distribution networks have been enormously extended, probably with less attention paid to the
optimum growth, to meet the rapidly growing demand. Distribution automation (DA) is
getting world wide attention to overcome these problems. Distribution automation systems
(DAS) can improve the operation of distribution system and the quality of supply. DA aims at
conservation of energy, including reduction of consumption and losses in the distribution and
transmission circuits, reduction of peak load, improvement in the reliability and quality of
service, defenal of new construction, and recovery of lost revenue.
Prominent features of electrical dish·ibution are radial or near radial structure; multiphase,
unbalanced, grounded or ungrounded operation; dispersed generation; multiphase, multi-mode
control distribution equipment; unbalanced distributed loads and extremely large number of
branches/nodes. In order to perform the desired functions of distribution management system
(DMS), a data acquisition system for distribution networks is needed similar to the real time
supervisory control and data acquisition system (SCADA) of energy management system
(EMS) used for transmission networks. Some of DA application functions are distribution
power flow, distribution state estimation, distribution short circuit analysis, dish·ibution fault
location, distribution feeder reconfiguration, and distribution feeder voltage and var control.
However, the on-line power flow and state estimation algorithms employed in EMS cannot be
used in a DMS. This is because the basic structure of the EMS on-line power flow and state
estimation is based on the assumptions which are not generally valid for distribution systems.
The transmission systems are generally assumed to operate under balanced three-phase
conditions, and the network is a symmetrical three-phase system that is fully described with its
positive sequence network. The RIX ratio of distribution lines in general are high and the fast
decoupled models and solution methods based on the assumption that RIX ratio of lines are
quite small, may fail to provide solution to the distribution networks. The distribution systems
may also require information of all the phasor quantities, as the system is generally
unbalanced.
In view of the above, this dissertation concentrates on the development of new and robust
algorithms related to network configuration, power flow analysis, reactive power
compensation, state estimation and fault analysis for application in distribution automation.
For reliable supply of power to consumers, distribution networks are fed from alternative
sources/substation feed-points. In real time environment distribution network configuration
changes dynamically due to switching. Real time network model depends on the conectness of
the network topology determined from the telemetered data. This dissertation presents a
network topology processing (NTP) algorithm suitable for distribution networks. A simple
data structure for distribution network connectivity information storage is proposed for
l1l
efficient implementation of network topology processing. The developed method has been
tested on a large distribution network with several feeders.
An efficient load flow solution technique is required as a part of the distribution automation
system for taking various control and operation decisions. A robust three-phase power flow
algorithm is presented in this dissertation. This method exploits the radial nature of the
network and uses forward and backward propagation technique to calculate branch currents
and node voltages. The proposed method considers all aspects of three-phase modelling of
branches and detailed load modelling. The merits of the method are, guaranteed convergence
even for heavily loaded network with poor voltage profile. The method has been tested on
practical distribution systems with many feeders emanating from grid substation with large
number of nodes and branches. The application of the proposed method is also extended to
find optimum location for reactive power compensation and network reconfiguration for
planning and day-to-day operation of distribution networks.
Distribution state estimators (DSE) will also play a critical role in distribution management
system to estimate those real-time system states which are unable to be obtained from the
limited measurement instruments in the distribution network. The success of DAS largely
depends on the availability of reliable database of the control centre and thus requires an
efficient state estimation (SE) solution technique. An efficient three-phase state estimation
algorithm for application to radial distribution networks is presented in this dissertation. The
method estimates the line flows, node voltages and loads at each node based on the measured
quantities. The SE cannot be executed without adequate number of measurements. The
extension of the method to the network observability analysis and bad data detection is also
discussed. The proposed method has been tested on a few sample and practical distribution
networks with simulated data for real-time measurements.
Unlike in transmission system, distribution networks may not be provided with protective
devices or circuit breakers in each branch of the feeder. Though, RTUs may be installed at
various nodes/branches of the feeder for various measurements, circuit breakers may be only
at the substation/switching station in the network. For any fault in the feeder, a large pait of
the feeder, may be isolated depending on the circuit breaker installation. For the purpose of
speedy repair work and maintenance, it is important to find the exact fault location and type of
fault. An algorithmic approach for finding the location and type of fault based on the three
phase measurements obtained for state estimation is presented. Results of the simulated fault
conditions on practical distribution systems are also presented.
Several 11 kV sample systems of 12 node, 18 node, 19node, 28 node, 38 node and a 132/33
kV practical system with 7 major feeders, nearly 1000 nodes, several switching stations/feedpoints
are used for testing the proposed algorithms in this dissertation. Results of the studies
indicate that the developed algorithms are suitable for application to practical systems. |