Kısa dönem elektrik talep tahminleri için yapay sinir ağları ve uzman sistemler tabanlı hibrit sistem geliştirilmesi

Elektrik enerjisi modern dünyada yüksek refah seviyesi ve rahat yaşam standartları açısından en fazlaönem taşıyan metalardan bir tanesidir. Elektrik arz güvenliğinin sağlanabilmesi doğru elektrik taleptahminlerinin yapılmasını gerektirmektedir. Bu çalışmada, Türkiye'nin piyasa ve mevsimsel koşulları gözönüne alınarak, yapay sinir ağları ve uzman sistemlerin birlikte kullanıldığı, kısa vadeli elektrik taleptahminlerinde yüksek doğrululuk derecesi sağlayan bir hibrit sistem geliştirilmiştir. EPSİM-NN adı verilen yeni tahmin sisteminde, günlük ortalama saatlik talep miktarı ve 24 saatlik talep şekli iki farklı yapay sinirağı kullanılarak belirlenmektedir. Bu ağlardan elde edilen sonuçlar birleştirilerek günlük talep tahmini eldeedilmektedir. Hesaplanan tahmin değerleri, yakın zaman talep trendlerinin kullanıldığı bir uzmansistemden geçirilerek tahminlerdeki hatalar minimize edilmektedir. Söz konusu sistem kullanılarak Türkiyeiçin yapılan tahminlerin gerçekleşen değerlerle karşılaştırılması sonucunda, EPSİM-NN tarafından oluşturulan tahminlerin doğruluk derecelerinin oldukça yüksek olduğu görülmüştür

Development of a hybrid system based on neural networks and expert systems for shortterm electricity demand forecasting

Electrical power is one of the most important commodities in terms of high levels of welfare andcomfortable living standards in the modern world. The provision of electricity supply security requiresaccurate electricity demand forecasts. In this study, a hybrid system using neural networks and expert systems has been developed considering Turkey's electricity market and the seasonal conditions in ordertoobtain short-term electricity demand forecasts with high degree of accuracy. The new forecast system,which is called EPSIM-NN, estimates daily average per hour demand and 24-hour shape function using two different artificial neural networks. The results from these two separate networks are combined toobtain 24-hour daily demand estimates. Forecast errors are further minimized by an expert system module using correction factors derived from recent demand data. By comparing the estimated values with theactual values for typical Turkish demand scenarios, we conclude that degree of accuracy is quite highforEPSIM-NN generated forecasts.

PDF

___

1. Liu N., Babushkin V., Afshari A., Short-Term Forecasting of Temperature Driven Electricity Load Using Time Series and Neural Network Model, Journal of Clean Energy Technologies, 2 (4), 327-331, 2014
2. Ericson T., Short-term electricity demand response, Thesis for the degree doctor ingeniør, Norwegian University of Science and Technology, Faculty of Information Technology, Trondheim, March, 2007.
3. Kandananond K., Forecasting Electricity Demand in Thailand with an Artificial Neural Network Approach, Energies 4 (1), 1246-1257, 2011.
4. AbuAl-Foul B., Forecasting Energy Demand in Jordan Using Artificial Neural Networks, Topics in Middle Eastern and African Economies, 14 (1), Sept 2012.
5. Kheirkhah A., Azadeh A., Saberi M., Azaron A., Shakouri H., Improved Estimation of Electricity Demand Function by Using of Artificial Neural Network, Principal Component Analysis and Data Envelopment Analysis, Computers &Industrial Engineering, 64 (1), 425-441, 2013.
6. Assareh E., Behrang M.A., Assareh R., Hedayat N., Integration of Artificial Neural Network and Intelligent optimization techniques on world electricity consumption estimation, World Academy of Science, Engineering and Technology, 73 (1), 2011.
7. Chogumaira E.N., Hiyama, T., Short-Term Electricity Price Forecasting Using a Combination of Neural Networks and Fuzzy Inference, Energy and Power Engineering, 3, 9-16, 2011.
8. Ozoh P., Abd-Rahman S., Labadin J., Apperley M., A Comparative Analysis of Techniques for Forecasting Electricity Consumption, Int. Journal of Computer Applications, 8(15), 2014.
9. Baliyan A., Gaurav K., Mishra S.K., A Review of Short Term Load Forecasting using Artificial Neural Network Models, Int. Conf. on Computer, Communication and Convergence (ICCC2015), 48(1), 121-125, 2015.
10. Baziar A., Kavousi-Fard A., Short Term Load Forecasting Using A Hybrid Model Based On Support Vector Regression, Int. Journal of Scıentıfıc & Technology Research, 4 (5), May 2015.
11. Es H.A., Kalender F.Y., Hamzaçebi C., Forecasting the net energy demand of Turkey by artificial neural networks, Journal of the Faculty of Engineering and Architecture of Gazi University, 29 (3), 495-504, 2014.
12. Soysal A., Akkurt A., Yapay Sinir Ağları ve Türkiye Elektrik Tüketimi Tahmin Modeli, Yüksek Lisans Tezi, İstanbul Teknik Üniversitesi, Fen Bilimleri Enstitüsü, 2005.
13. Aslan Y., Yaşar C., Nalbant A.,Electrical Peak Load Forecasting in Kütahya with Artificial Neural Networks, Dumlupınar Üniversitesi, Fen Bilimleri Enstitüsü Dergisi, 11 (1), 2006.
14. Enerji Piyasaları İşletme A.Ş. (EPİAŞ), seffaflik.epias.com.tr/transparency/tuketim/tahmin/yuk -tahmin-plani.xhtml
15. Philip C., Jackson Jr., Introduction to Artificial Intelligence, Dover Publications, 2. edition, June 1985.
16. Samarasinghe S., Neural Networks for Applied Sciences and Engineering: From Fundamentals to Complex Pattern Recognition, Auerbach Publications, First Edition, Sept 2006.
17. Otkun Ö., Doğan R.Ö., Akpınar A.S., Neural Network Based Scalar Speed Control of Lınear Permanent Magnet Synchronous Motor, Journal of the Faculty of Engineering and Architecture of Gazi University, 30 (1), 395-404, 2015.
18. Giarratan J., Expert Systems: Principles and Programming, Course Technology, 4. Edition, Oct 2004.
19. Yurtcu1 Ş., Özocak A., Prediction Of Compression İndex Of Fine-Grained Soils Using Statistical And Artificial İntelligence Methods, Journal of the Faculty of Engineering and Architecture of Gazi University, 31 (3), 597-608, 2016.
20. Yük Tevzi Bilgi Sistemi, ytbs.teias.gov.tr