Forecasting energy use and efficiency in transportation: Predictive scenarios from ANN models

Abstract

Transport energy efficiency measurements are a challenge for national-level organizations in developing nations. For this reason, a transportation method using 40 artificial neural networks (ANN) has been created to address this challenge. A model has been developed to forecast energy efficiency measurements using social and economic factors based on data collected from 28 European nations. After that, we compare the assumed data with the overall energy use using a bottom-up technique. Since Morocco does not have any energy efficiency parameters, it is utilized as an instance study. Energy demands were calculated at a highly disaggregated level, proving the model's remarkable performance. By providing the model with a number of different assumptions, we were able to predict energy usage up to the year 2050. The four potential future states of energy consumption were as follows: modal shift, vehicle electrification, frozen efficiency, and the application of EU regulations. Results from predictive scenarios state: (1) a 75% rise in energy demand due to frozen efficiency by 2050; (2) a 30% reduction in consumption due to the European Union fuel regulations; (3) a demand reduction to 7.3 Mtoe due to vehicle electrification; and (4) a 62% reduction in energy use and an 80% reduction in emissions due to mode shifts. Redistributing passenger miles and tonne-kilometers to increase average capacity and average load showed a greater likelihood of saving energy. One potential option to reduce greenhouse gas emissions was to replace diesel with biofuel for smaller vehicles and buses. The created model provides decision-making organizations with the resources required to figure out critical issues, execute policies, and reallocate infrastructure.