A Method Based on Machine Learning Techniques for the Development of a Parametric Environmental Impact Model for Industrial Electric Vehicles

Designers need to be aware early in the design phase of the environmental impact of their choices over the entire product life cycle. This paper proposes an eco-design method to support designers of different categories of electric vehicles, such as self-driving vehicles, cars, shuttles and buses. The methodology developed aims to realize a model for predicting the environmental impact of industrial electric vehicles. The proposed approach exploits machine learning methods to develop models with the design features of a generic electric vehicle, such as vehicle mass and distance traveled during its entire lifetime as independent parameters, to estimate the emissions of new products. The environmental impact indicator for this study is Climate Change, the dependent parameter chosen for the impact model. Machine learning algorithms were trained on training data retrieved from an automatic environmental impact estimation software tool based on an analytical approach. All stages of the product life cycle have been considered in the construction of the database, and the model provides quantitative results that consider the consumption of material and energy resources. Finally, the model is tested by estimating the environmental impact of a tourist shuttle.