Politecnico di Torino (logo)

Deep learning solution for the performance analysis of hybrid powertrains

Simone Quaranta

Deep learning solution for the performance analysis of hybrid powertrains.

Rel. Daniela Anna Misul. Politecnico di Torino, Corso di laurea magistrale in Automotive Engineering (Ingegneria Dell'Autoveicolo), 2021

PDF (Tesi_di_laurea) - Tesi
Licenza: Creative Commons Attribution Non-commercial No Derivatives.

Download (3MB) | Preview

Over the past years, the engine emission regulation has become of crucial importance and the development of Hybrid Electric Vehicles (HEV) is one of the best solutions in the market. Besides the emissions reduction, this class of vehicles has to satisfy important performances constraints. For this reason, the aim of this project is to develop a new algorithm capable of classifying the admissibility of hybrid electric vehicle layouts, which means to check if the performance index of each layout is lower than a predefined threshold. This algorithm exploits Deep Neural Network (DNNs), an artificial neural network (ANN) with multiple layers between the input and output layers. There are different types of neural networks, but they always consist of the same elements: neurons, synapses, weights, biases, and functions, used to replicate the human brains’ behaviour. This structure enables the algorithm to train and learn by itself. This admissibility classifier has been inserted into an existing pipeline, creating a new one composed of three different consecutive nets: a feasibility classifier, an admissibility classifier and a regressor of CO2. Each layout passes through this pipeline to verify if it satisfies the constraints of emissions (feasibility) and performance (admissibility) and, then, to predict its CO2 emissions. For this project, Spyder Anaconda has been used: it is an open-source cross-platform integrated development environment (IDE) for scientific programming in the Python language and libraries as Keras and Tensorflow.

Relators: Daniela Anna Misul
Academic year: 2020/21
Publication type: Electronic
Number of Pages: 66
Corso di laurea: Corso di laurea magistrale in Automotive Engineering (Ingegneria Dell'Autoveicolo)
Classe di laurea: New organization > Master science > LM-33 - MECHANICAL ENGINEERING
Aziende collaboratrici: UNSPECIFIED
URI: http://webthesis.biblio.polito.it/id/eprint/17497
Modify record (reserved for operators) Modify record (reserved for operators)