Summary: | 碩士 === 國立臺灣海洋大學 === 電機工程學系 === 103 === After the electricity act bill went through in recent years, the Taiwan Power Company (TPC) has been planning to initiate a new reform about electricity liberalization. In liberalized-electricity market, frequency regulation and spinning reverse, both of which are indispensable in the process of security-constrained dispatch, are provided by ancillary-service market. All participants in the market have to assume the cost of security dispatch system. The safety concerns can arise if the services of frequency regulation and spinning reverse are inadequate; on the contrary, unnecessary costs will make the system inefficient if the services are too much. Therefore, it is beneficial for electric power industry to develop calculation methods to make balances between frequency regulation and spinning reverse.
This thesis starts with discussions on system dynamics when isolated power systemsmalfunction by means of comparing calculation methods of different countries, and use Statistical Method and Neural Network Method, combining with historical datas, to estimate the demand of frequency regulation and spinning reverse, then predict the demand of frequency regulation and spinning reverse as system, season, or time interval varies,then simulate and verify the demand in PSS/E.Finally, this paper expect to establish an estimate mechanism for frequency regulation andspinning reverse forecast as a basis for the liberalization of TPC, or other regional industry.
|