Abstract:
Transmission and Distribution losses in India are much higher than those in any developed country. Out of these, the losses in distribution sector are very large. Commercial losses constitute the major part of losses in the distribution system. Estimation of these losses is of major importance which requires optimal placement of metering devices. In this thesis, the research work is done in the area of this meter placement problem. A 41-bus 33 kV system is considered for the analysis in the entire thesis.
In the present thesis work, the meter placement problem is handled using genetic algorithms. The meter placement thus obtained, balances between the financial constraint and Observability of the network. Numerical Observability technique is used to find the Observability of the network for each and every combination of meters obtained for meter placement.
The meter placement problem is found to have some disadvantages, hence a new algorithm is developed which overcomes the disadvantages of the former. This algorithm divides the basic structure of the network in to number of sub-laterals, processes them separately and again combine them to form the original structure. More constraints are imposed to refine the meter structure. The original method and the proposed method are compared in the aspect of faster convergence, search space, cost of the algorithms and size of matrices used.
The meters thus obtained will monitor the different areas of the network. The commercial losses are estimated with the help of these monitoring meters. The commercials losses and Technical losses of the 41-bus, 33 KV system are compared. Load flow analysis is done by Forward/Backward sweep method.
