Nodal load estimation for electric power distribution systems

Up-to-date nodal load information is required to accurately represent customers in electric power distribution systems. In addition, many operating, control and planning studies in power distribution systems directly depend on nodal load information. Thus, considering the load driven nature of power systems and the interdependence of loads and system states (voltage magnitudes and phase angles), this thesis defines and studies nodal load estimation in power distribution systems with limited synchronized measurements.; This thesis presents a constrained, nonlinear optimization problem formulation that (i) treats loads as varying parameters, (ii) incorporates real-time measurements and (iii) includes practical guidelines issued by operating and loading constraints. Locally convergent algorithms are designed for power distribution networks of different structure. For radial networks, sensitivity-based network decomposition methods are presented. For general structure networks, two weighted least squares (WLS) approaches are designed. Extensive simulation results on a 20-bus unbalanced, radial test system and an actual 394-bus unbalanced distribution network with radial or meshed structures are obtained. Finally, the meter placement problem in power distribution systems is investigated for enhanced load estimation. A heuristic incremental method is presented with promising results.