摘要 本研究應用ARMA 階數模型預測「台電供電處-新伯公倉庫斜屋頂太陽能電廠」發電量。一般利用AR或MA配適模型時,有時需要較多參數才能配適出合適的模型,為了克服此點困難,利用ARMA模型或許可以精簡配適模型之參數。研究結果顯示,預測時的RMSE介於9.2449至12.8915之間。對於R值而言,預測時的R值介於0.39至0.69之間。以ARMA (2, ) 對新伯公倉庫斜屋頂太陽能電廠發電量進行預測時,以ARMA (2,9) 模型之結果較佳,在預測發電量時,RMSE為10.0013,R值為0.63。
Abstract This study used an ARMA stage model to predict electrical production of “Taipower Power Station of Xinbogong Warehouse slant roof solar power plant”. Utilizing either AR or MA to fit our models would sometimes require more variables. In order to overcome this limitation, using an ARMA model may simplify and reduce the variables required to fit our model. The results showed that predicted RMSE was between 9.2449 and 12.8915, with an R value between 0.39 and 0.69 When predicting the electrical production of Xinbogong Warehouse Slant-roof Solar Power Plant, this study obtained better results using an ARMA (2,9) model, with an RMSE was 10.0013 and R value was 0.63.