Seite - 74 - in Intelligent Environments 2019 - Workshop Proceedings of the 15th International Conference on Intelligent Environments

can be incorporated easily to electricity load models using data from port and airport control points. This chapter has investigated these special circumstances studying Balearic Islands as a particular case and forecasting for lead times from 1 to 10 days ahead, which coincide with the forecasting system currently implemented at Red Eléctrica de España, the Spanish system operator. Using simple multiple regression and time series models, results have evidenced again the importance of including meteorological variables in daily electricity forecasting. However, the most significant result has been the usefulness of including daily population measures that is found to be relevant in improving the forecasting accuracy. Thus, the results have shown that the major improvement in error reduction comes from understanding how the load reacts to population stock variable, and integrating its effects together with the weather variables and other specific dummies in an extended model that captures the main determinants of the electricity load. In general, and depending on the particularity of the islands, the inclusion of either HPDI variable or airport’s arrival variable improves the forecasting performance of the dynamic model ARMAX. The use of HPDI variable in the dynamic model in the case of Majorca and Pitiüses has shown a very good forecasting performances in annual average and in high season. Finally, in the case of Minorca the dynamic models that incorporate airport’s arrival variable perform better in all seasons and in annual average compared to their correspondent that involves HPDI variable. Inclusion of tourist variables in forecasting electricity models can be of enormous interest for tourism regions to reduce the risk of load shedding and power blackout. However further research would have to consider the inclusion of such variables by using more advanced techniques such as periodic autoregressive models. In the end, there is a strong potential for the use of population stock variable in improving the accuracy and uncertainty assessment of electricity demand forecasts for number of tourist destinations with the same characteristics as Balearics Islands. References [1] M. Bakhat, J. Rosselló, O. Sáenz-de-Miera, Developing a daily indicator for evaluating the impacts of tourism in isolated regions, European Journal of Tourism Research 3 (2010), 114-118. [2] J.R. Cancelo, A. Espasa, Modeling and forecasting daily series of electricity demand, Investigaciones Económicas, 20 (1996), 359-376. [3] J.R. Cancelo, A. Espasa, R. Grafe, Forecasting the electricity load from one day to one week ahead for the Spanish system operator, International Journal of Forecasting 24 (2008), 588-602. [4] T.J. Considine, The impacts of weather variations on energy demand and carbon emissions, Resource and Energy Economics 22 (2000), 295–314. [5] A. Da Silva, V. Ferreira, R. Velasquez, Input space to neural network based load forecasting, International Journal of Forecasting 24 (2008), 616-629. [6] C. Deb, F. Zhang, J. Yang, S. E. Lee, and K. W. Shah, A review on time series forecasting techniques for building energy consumption, Renewable and Sustainable Energy Reviews, 74 (2017), 902–924. [7] I. Dobson, B.A. Carreras, D.E. Newman, A probabilistic loading-dependent model of cascading failure and possible implications for blackouts, 36th Annual Hawaii International Conference on System Sciences (HICSS'03) - Track 2, Big Island, Hawaii,2003, p. 65a. [8] V. Dordonnat, S.J. Koopman, M. Ooms, A. Dessertaine, J. Collet, An hourly periodic state space model for modelling French national electricity load, International Journal of Forecasting 24 (2008), 566-587. [9] R.F Engle, C. Mustafa, J. Rice, Modeling peak electricity demand, Journal of Forecasting 11(1992), 241- 251. [10] E. Gabreyohannes, A nonlinear approach to modelling the residential electricity consumption in Ethiopia, Energy Economics 32 (2010), 515-523. M.BakhatandJ.RosselloNadal / ImprovingDailyElectricityLoadsForecasting74