Electrical load forecasting using Hijri causal events
Download as PPT, PDF0 likes145 views
This document discusses the enhancement of electrical load forecasting by considering the impact of Hijri causal events, particularly during Ramadan and religious holidays. Two models using artificial neural networks (ANN) were applied, with one incorporating Fourier series to account for seasonal patterns, resulting in reduced forecasting errors. The findings suggest significant improvements in forecasting accuracy when hijri calendar effects are included.
Electrical load forecasting using Hijri causal events
- 1. Electrical Load Forecasting Using Hijri Causal Events Elsayed E. Hemayed and Maged M. Eljazzar Computer Engineering Dept. Faculty of Engineering Cairo University, Egypt mmjazzar@ieee.org 2016 Eighteenth International Middle-East Power Systems Conference (MEPCON) December 27-29, 2016 - Helwan University, Cairo – Egypt 1
- 2. Outline – Introduction – Objective – Previous work – Hijri calander – Model – Experimental Results – Conclusions and future work 2
- 3. Introduction – Why Load forecasting is important ? – Types of load forecasting. – Machine learning techniques (ANN, SVM). – Statistical techniques (ARIMA, regression). – Load forecasting parameters. – Casual events. 3
- 4. Objective – Our goal is to study the effect of Ramadan and religious holidays to match the consumers behavior for better forecasting. 4
- 5. Literature review – Meteorological parameters and Special days effect. – The effect of holidays not only influences special days, but it also influences the previous and next days. 5
- 6. Hijri calendrer – Two main issues considered in special days:- • working hours • people activities – Besides the effect of the day and its impact on electric demand, and the temperature. 6
- 7. The hourly peak load of a day in Ramadan vs. a normal day 7
- 8. The daily load profile 8
- 9. Model – Two models are applied model 1 using ANN and model 2 Using ANN with Fourier series to consider • Monthly and week effects. • Detecting seasonal pattern. • The accumulative effect of special days. – Two different architecture are introduced. 9
- 10. Residuals 10
- 11. Model – The red line represents the forecasting without considering the effect of casual Hijri events. The blue line gave better results than the red line. – The effect of Ramadan (22nd of August to 20th of Sept.) is very clear near the end of the year. 11
- 12. Results Parameter ANN model 1 ANN model 2 RMSE(GW/H) 9.429274 0.9634765 MAE (GW/H) 5.882814 0.6359111 MAPE (%) 0.2660583 0.02835947 12
- 13. Conclusions – In this paper, we presented the main Hijri casual events during the Islamic calendar. – We applied ANN model before and after including those Hijri casual events. Using the Hijri calendar casual events, with ANN model, for load forecasting reduced the forecasting errors. 13
- 14. Future work – Using more recent data. – Applying on short term load forecasting. – Using different scenarios. 14
- 15. Thank you for further questions mmjazzar@ieee.org 15