Voltage Dip (Sag) Estimation in Power Systems based on Stochastic Assessment and Optimal Monitoring

Doktorsavhandling, 2005

This dissertation deals with the statistical characterization of the performance of power systems in terms of voltage dips (sags). It presents a method named voltage dip estimation that extends monitoring results to buses not being monitored. Statistical dip characterization of power networks is essential to decide about mitigation methods as well as for regulatory purposes. Statistics on voltage dip may be obtained by means of 1) monitoring of the power supply, and/or 2) stochastic assessment of voltage dips. Monitoring is expensive and requires long monitoring periods. Stochastic assessment is a simulation method that combines stochastic data concerning the fault likelihood with deterministic data regarding the residual voltages during the occurrence of faults. The method of fault positions is used in this work to assess the dip performance of an existing power system. An alternative analytic approach to the method of fault positions that addresses the question of the suitable number of fault positions is proposed. It is shown that despite being able to provide a description of the long-term expected performance of the network, the method of fault positions cannot predict the performance during a particular year. Comparisons of pseudo measurements with the prediction via the method of fault positions show discrepancies and therefore the need for adjustment. A Monte Carlo simulation approach is proposed to better describe the performance of the network, and an optimal monitoring program is suggested to accomplish the adjustment of the method of fault positions. An integer optimisation model is introduced in order to determine the optimal number and location of monitors, so that every fault triggers at least a given number of power quality meters. The results of the monitoring program are then used to perform voltage dip estimation, adjusting the stochastic assessment and extending monitoring results to non-monitored buses. The dissertation shows that it is possible to profile the dip performance of the entire power system without the need for power quality monitors at every load bus of the power network.