Some Aspects of Static and Dynamic Distribution System State Estimation with Optimal Meter Placement Studies

Distribution System;State Estimation;Optimal Meter Placement;Iterative Extended Kalman Filter

Kandenkavil, Sreedevi Valsanadvisor:Bhattacharya, Kankar ; affiliation1:Faculty of Engineering ; Bhattacharya, Kankar ;

University of Waterloo

关键词: Distribution System;    Optimal Meter Placement;    State Estimation;    Master Thesis;    Iterative Extended Kalman Filter;   

Others  :  https://uwspace.uwaterloo.ca/bitstream/10012/13728/3/Kandenkavil_SreedeviValsan.pdf

瑞士|英语

来源: UWSPACE Waterloo Institutional Repository

【 摘 要 】

In a power distribution system, due to the evolution of Active Distribution Networks(ADNs), there is a possibility of violation of the system operational constraints. A stateestimator provides an approximate snapshot of the distribution system operation when thebus voltages and power measurements are available. Thus it plays a key role in monitoringthe system, thereby ensuring a safe state of operation. According to the nature of thesystem, Distribution System State Estimation (DSSE) can be classified in to static DSSEand dynamic DSSE. Static DSSE is commonly designed as a Weighted Least Square (WLS)estimator using either bus voltages or branch currents as system states. For dynamic DSSE,the performance of static state estimators are limited. A Kalman filter based state estimatorcan be used in such time varying systems. A study of the algorithms used for these twoDSSE methods is necessary in order to analyze the factors affecting the estimation accuracy.In a power distribution system, with limited availability of measurements, and additionalmeasurements being expensive, careful selection of the location for the placement of metersbecomes important. The measurement meters typically considered are Phasor MeasurementUnits (PMUs) and power (PQ) meters. The existing placement problems lay more emphasison minimizing the cost of installing such meters, while the quality of estimation remainsignored. Thus there is a need to formulate methods for optimal allocation of meters in acost effective way without altering the accuracy of DSSE.In this work, a detailed study is conducted on the two static DSSE algorithms, Node Voltagebased State Estimation (NVSE) and Branch Current based State Estimation (BCSE)and the DSSE performance is compared based on Average Root Mean Square (ARMSE)Value of state estimates. The thesis also analyzes the impact of the number of PMU measurementsavailable on DSSE performance. Several optimization based approaches are proposedto address the optimal meter placement problem considering different objectives suchas minimization of cost, WLS residual estimate, a multi-objective function comprising costand WLS, and the ARMSE of the estimated bus voltage. An Iterative Extended KalmanFilter (IEKF) is used for performing dynamic DSSE. The dependency of various parameterssuch as selection of time frame, apriori estimate information length and PMU measurementerrors on the accuracy acquired by DSSE is also presented.The studies and proposed models are simulated in a 33-bus distribution feeder. Theresults illustrating the efficiency and speed of convergence of different static and dynamicDSSE methods are discussed. The various optimization models for meter allocation areformulated and compared based on meter placement cost and ARMSE of voltage estimates.