Modeling and forecasting of Turkey’s energy consumption using socio-economic and demographic variables Murat Kankal a , Adem Akpınar a,b,⇑ , Murat _ Ihsan Kömürcü a , Talat S ßükrü Özs ßahin a a Karadeniz Technical University, Civil Engineering Department, 61080 Trabzon, Turkey b Gümüs ßhane University, Civil Engineering Department, 29000 Gümüs ßhane, Turkey article info Article history: Received 24 May 2010 Received in revised form 11 November 2010 Accepted 1 December 2010 Available online xxxx Keywords: Energy consumption Artificial neural network Regression models Turkey abstract This study deals with the modeling of the energy consumption in Turkey in order to forecast future pro- jections based on socio-economic and demographic variables (gross domestic product-GDP, population, import and export amounts, and employment) using artificial neural network (ANN) and regression anal- yses. For this purpose, four diverse models including different indicators were used in the analyses. As the result of the analyses, this research proposes Model 2 as a suitable ANN model (having four independent variables being GDP, population, the amount of import and export) to efficiently estimate the energy con- sumption for Turkey. The proposed model predicted the energy consumption better than the regression models and the other three ANN models. Thus, the future energy consumption of Turkey is calculated by means of this model under different scenarios. The predicted forecast results by ANN were compared with the official forecasts. Finally, it was concluded that all the scenarios that were analyzed gave lower estimates of the energy consumption than the MENR projections and these scenarios also showed that the future energy consumption of Turkey would vary between 117.0 and 175.4 Mtoe in 2014. Ó 2010 Elsevier Ltd. All rights reserved. 1. Introduction The increasing worldwide demand for energy requires develop- ment of intelligent forecasting methods and algorithms. The esti- mation of energy demand based on economic and non-economic indicators may be achieved by certain linear or non-linear statisti- cal, mathematical, and simulation models. The non-linearity of these indicators and energy demand has led to a search for intelli- gent solution approach methods such as genetic algorithms, fuzzy regression, and neural networks. ANNs have been used in non- linear modeling and forecasting [1]. Modeling of energy consumption is usually based on historical consumption and the relationship of this consumption with other relevant variables, such as economic, demographic, climatic indica- tors, and such [2]. At present, energy modeling is a subject of wide- spread interest among engineers and scientists concerned with the problems of energy production and consumption [3]. Modeling in some areas of application is now capable of making useful contri- butions to planning and policy formulation [4]. In this regard, energy planning is not possible without a reasonable knowledge of the past and present energy-consumptions and likely future de- mands [5]. Modeling and prediction of energy consumption play a vital role in developed and developing countries for policy makers and related organizations. Underestimation of the consumption would lead to potential outages that are devastating to life and economy, whereas overestimation would lead to unnecessary idle capacity which means wasted financial resources. Therefore, it would be better to model electricity energy consumption with good accuracy in order to avoid costly mistakes. Also, it is better to use models that can handle nonlinearities among variables as the expected nature of the energy consumption data is non-linear [6]. Traditionally, regression analysis has been the most popular modeling technique in predicting energy consumption. However, the importance of the ANN approach, apart from reducing the time required, is that it is possible to make energy applications more viable and thus more attractive to potential users, such as energy engineers. Also, this approach has the advantages of computational speed, low cost for feasibility, and ease of design by operators with little technical experience. Therefore, the use of ANN for modeling and prediction purposes is becoming increasingly popular in the recent decades. This is mainly because ANN has very good approx- imation capabilities and offers additional advantages, such as short development and fast processing times. ANNs are especially useful in predicting problems where mathematical formulae and prior knowledge on the relationship between inputs and outputs are un- known [5,7–9]. The most important scope of this study is to present a new and different model, which is not given in the literature, for energy 0306-2619/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.apenergy.2010.12.005 ⇑ Corresponding author. Tel.: +90 456 233 7425x242; fax: +90 456 233 7427. E-mail addresses: aakpinar@ktu.edu.tr, ademakpinar@hotmail.com (A. Akpınar). Applied Energy xxx (2010) xxx–xxx Contents lists available at ScienceDirect Applied Energy journal homepage: www.elsevier.com/locate/apenergy Please cite this article in press as: Kankal M et al. Modeling and forecasting of Turkey’s energy consumption using socio-economic and demographic variables. Appl Energy (2010), doi:10.1016/j.apenergy.2010.12.005