Abstract
Deregulation of electric power industry has become a worldwide trend, undergoing in many countries including Taiwan. Because of the liberalization policy of Taiwan’s power industry, there is a critical demand to explore the mechanism of power market and thus design appropriate market structure to maintain the stability of market in a deregulated environment. However, given the characteristics of the Taiwanese power industry, specifically limited transmission capability and insufficient stability, the transmission capacity constraints are important for operating and planning power systems. Furthermore, distribution feeder fault causes power outages and will also significantly affect power systems’ reliability, security and quality, among other important factors. Therefore, it is important to apply useful methods to diagnose and thus locate the fault quickly to reduce the outage duration and avoid huge economical loss. In practice, feeder patrols in Taiwan Power Company (Taipower) usually identify the fault locations by referencing the regional distribution of the trouble calls, the abnormal observations of the feeders, and the observed conditions in the surrounding environments. Then, feeder patrols have to rush into the field along feeder to locate the fault mainly by visual inspection and by trial energization of the feeder, section by section. Such a trial feeder energization is harmful to the cable and frequently takes a long time for power restoration. This dissertation aims to explore the mechanism for power market by using scenario analysis for examining the locational prices with load variation and different bidding strategies and to thus propose the appropriate operation guides for avoiding possible failure of market operation in a deregulated electric power market. This dissertation discusses critical success factors of the proposed market mechanism. In addition, this dissertation also aims to develop a data mining framework based on rough set theory and association rule to derive useful patterns and rules for distribution feeder fault equipment diagnosis for fault location. In particular, the historical data of distribution feeder faults, which occurred within the business area of Taipei City District Office of Taipower was used for validation. The results have demonstrated practical viability of data mining approach for fault location based on historical data. The feeder patrols can locate the fault location and find the fault equipment quickly through the derived inference rules.