Applied Self-Reunion Multiple Regression Model to Short-Term Forecasting

• Main Authors: Heng-Chi Lin, 林宏基
• Other Authors: S.R. Huang
• Format: Others
• Language: zh-TW
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/y93vnq

碩士 === 逢甲大學 === 電機工程所 === 91 === Abstract An accurate short-term load forecast is an essential component of any Energy Management System (EMS). This short-term load forecast can be used to forecasts of either total load requirements (MWH) during a period of MW or peak loads requirements for the period. And based on some useable data, system dispatchers and operation system analysis are able to control and to plan power system. Further to accurate the varying nature of the load. This paper presents a method of forecasting the hour (day) load demand on a power system. The method of forecasting uses self-reunion (S.R.) analysis nonlinear multiple regression models with linear planning method to solve the forecasting parameters. With the main proposed threshold models algorithm, we can use fewer parameters to capture the random component in dynamics load to estimate the load forecasting parameters. The results, based on Taiwan Power company (TPC) historical load demand, indicate that the proposed algorithm is capable of providing more accurate load forecast and on-line forecast. Keywords: Nonlinear、Multiple-Regression、Self-Reunion、Short-Term Load Forecast