:::

臺灣能源期刊論文全文

臺灣能源期刊第6卷第3期內容

出刊日期:September, 2019

題目
整合自我迴歸移動平均與類神經網路於電力負載之預測改善
Title
Forecast Improvement of Electricity Load Using a Model Integrating ARIMA and ANN
作者
陳春志、游凱為、許倢歆、楊世銘
Authors
C. J. Chen, K. W. Yu, C. H. Hsu, S. M. Yang
摘要
本文發展一整合季節性自我迴歸移動平均(SARIMA)和類神經網路(ANN)演算法之預測模型,利用歷史電力負載數據、氣象數據和假日效應之變量當作輸入參數,模擬電力系統動態及電能供應穩定性之預測。整合的SARIMA-ANN方法可用於預測顯著的季節性及周期性特徵之電力負載系統數據。研究模擬結果顯示,此模型應用在預測能力方面比ANN模型、ARIMA模型、SARIMA模型和ARIMA-ANN模型更有效,藉由使用此預測模型可減少技術性因子的影響並能產生更好的預測結果。
關鍵字
類神經網路,自我迴歸移動平均,短期負載預測
Abatract
A model combining seasonal autoregressive integrated moving average (SARIMA) and artificial neural network (ANN) methodology is developed to simulate the dynamics and to forecast stable electrical energy supply in power system by using the input of historical daily electricity load data, weather data, and holiday effect variables. An integrated SARIMA-ANN method is to process the strong seasonality and periodic characteristics of load data. Simulation results show that the proposed model is more effective than ANN model, ARIMA model, SARIMA model and ARIMA-ANN model in prediction and forecasting. The technical factors are attenuated by using the model to yield better forecasting results.
Keywords
Artificial Neural Network, Autoregressive Integrated Moving Average, Short-term load forecasting