اخبار و اعلانات

 آمار

تعداد نشریات28
تعداد شماره‌ها874
تعداد مقالات8,421
تعداد مشاهده مقاله17,113,369
تعداد دریافت فایل اصل مقاله11,474,818
Keypour, Reza, Shirzadiraj, Vahid. (1405). An Improved Load Forecasting Approach Using LSTM Networks Enhanced by Modeled Demand Response Behavior. نشریه علمی دانشگاه سیستان و بلوچستان, (), -. doi: 10.22111/ieco.2026.55110.1755
Reza Keypour; Vahid Shirzadiraj. "An Improved Load Forecasting Approach Using LSTM Networks Enhanced by Modeled Demand Response Behavior". نشریه علمی دانشگاه سیستان و بلوچستان, , , 1405, -. doi: 10.22111/ieco.2026.55110.1755
Keypour, Reza, Shirzadiraj, Vahid. (1405). 'An Improved Load Forecasting Approach Using LSTM Networks Enhanced by Modeled Demand Response Behavior', نشریه علمی دانشگاه سیستان و بلوچستان, (), pp. -. doi: 10.22111/ieco.2026.55110.1755
Keypour, Reza, Shirzadiraj, Vahid. An Improved Load Forecasting Approach Using LSTM Networks Enhanced by Modeled Demand Response Behavior. نشریه علمی دانشگاه سیستان و بلوچستان, 1405; (): -. doi: 10.22111/ieco.2026.55110.1755

An Improved Load Forecasting Approach Using LSTM Networks Enhanced by Modeled Demand Response Behavior

International Journal of Industrial Electronics Control and Optimization
مقالات آماده انتشار، پذیرفته شده، انتشار آنلاین از تاریخ 02 آبان 1405    اصل مقاله (595.4 K)
نوع مقاله: Research Articles
شناسه دیجیتال (DOI): 10.22111/ieco.2026.55110.1755
نویسندگان
Reza Keypour* 1؛ Vahid Shirzadiraj2
1Faculty of Electrical and Computer Engineering, Semnan University, Iran
2Faculty of Electrical and Computer Engineering, Semnan University, Semnan, Iran
چکیده
The increasing penetration of renewable energy sources and the expansion of demand response (DR) programs have introduced considerable uncertainty into electricity consumption patterns, thereby reducing the accuracy of conventional load forecasting models. Most existing forecasting approaches model system load as a single aggregated variable and consequently lack the ability to disentangle and analyze the structural impacts of DR. This paper proposes a load forecasting framework that explicitly and interpretably incorporates DR into the forecasting process. To this end, an index‑based DR model (IDRM) is developed to quantify hourly DR using three fundamental indicators—available flexibility, willingness to participate, and participation level—providing a structured and operational representation of DR behavior. The extracted DR component is employed to decompose the observed load and estimate the baseline load, which is then forecasted using an LSTM neural network. The final load forecast is obtained by combining the predicted baseline with the reconstructed DR component. Experimental results based on real-world data demonstrate that the proposed LSTM-IDRM framework reduces the RMSE value by approximately 18.5% (from 0.395 to 0.322) compared with a conventional LSTM model, highlighting the importance of explicit DR modeling in enhancing load forecasting performance.
کلیدواژه‌ها
Load forecasting؛ load flexibility؛ baseline load estimation؛ Index based Demand Response Model (IDRM)؛ LSTM neural networks
آمار
تعداد مشاهده مقاله: 2


سامانه مدیریت نشریات علمی. طراحی و پیاده سازی از سیناوب