State and parameter estimation of power systems using phasor measurement units as bilinear system model

In the wake of past various blackouts, management and control of power system is going through essential transitions. To transform the modern grid into smart grid, various actions related to data acquisition and processing, security monitoring and control and energy/economy decisions have to be advanced to next level. Including of Phasor Measurement Units (PMUs) along with current SCADA (Supervisory Control and Data Acquisition) system is an essential step towards realization of smart grids.
PMUs can give synchronized measurements of voltage and current phasors from the grid, however their placement and maintenance could be daunting task based on their cost and the size of the grid. Hence various optimization techniques have been used to solve Optimal PMU placement (OPP) problem that provides complete observability of the grid. It was assumed that a PMU used in power system will have a very large number of channels facilitating current measurements of all the incident branches possible when placed on a node. This thesis solves a more practical problem when number of channels in the PMUs is comparatively smaller and provides a generalized formulation of OPP problem. The formulated problem is extended to various cases of OPP problem based on requirement of observability, cost ratio, presence of zero injection buses, requirement of measurement redundancy, presence of SCADA, combination of different types of PMUs etc.
For the power system network, matrices to define nodal connectivity and branch selectivity have been introduced. These matrices are used to formulate the generalized OPP problem for limited channel capacity PMUs. The problem has been tested on various standard IEEE test bus systems. Furthermore, application of PMU measurements for utilities must be addressed. This thesis develops a new model of state and parameter estimation of power systems using PMU measurements. The approach elaborates on bilinear model of power systems based on PMU measurements. Various correlations between states, parameters and PMU measurements have been included to make the solution more applicable to utility companies. A formula for parallel Kalman filter has been developed and implemented to do the state and parameter estimation of power systems. Parallel Kalman filter (PKF) is a special type of dynamic state estimation technique where a nonlinear model can be partitioned into two dependent linear models. Derivation of PKF is based on game theory where each opponent makes their move dependent on the outcome of the move played by the other opponent. Because of the knowledge of dynamics of the system and correlation between various errors along with errors due to approximation in parameters of the system, the developed dynamic state estimation technique tends to be more accurate than a dynamic state estimation technique using Kalman Filter, which in turn is more accurate than current static state estimation techniques.