@phdthesis{01HVRK4GCW2PWEXP2CPDGX72Q3,
  abstract     = {{The automotive industry is required to accelerate the development and deployment of electrified vehicles at a faster pace than ever, to align the transportation sector with the climate goals. Reducing the development time of electric vehicles becomes an urgent priority. On the other hand, the industry is challenged by the increasing complexity and large design space of the emerging electrified powertrains. The existing approaches to address component design, such as numerical methods exemplified by finite element method, computational fluid dy-namic, etc., are based on a detailed design process. This leads to a long computational burden when trying to incorporate them at system-level. Speeding up the early development phases of electrified vehicles necessitates new methodologies and tools, supporting the exploration of the system-level design space. These methodologies should allow for assessing different sizing choices of electrified powertrains in the early development phases, both efficiently in terms of computational time and with reliable results in terms of energy consumption at system-level. To address this challenge, this Ph.D. thesis aims to develop a scaling methodology for electric axles, allowing system-level investigation of different power-rated electric vehicles. The elec-tric axle considered in this thesis comprises a voltage source inverter, an electric machine, a gearbox, and a control unit. The scaling procedure is aimed at predicting the data of a newly defined design of a given component with different specifications based on a reference design, without redoing time and effort-consuming steps. For this purpose, different derivations of scaling laws of the electric axle components are thoroughly discussed and compared at com-ponent-level in terms of power loss scaling. A particular emphasis is placed on examining the linear losses-to-power scaling method, which is widely employed in system-level studies. This is because, this method presents questionable assumptions, and has not been the subject of a comprehensive examination. A key contribution of the presented work is the derivation of power loss scaling laws of gearboxes, which has been identified as a gap in the current litera-ture. This is achieved through an intensive experimental campaign using commercial gear-boxes. To incorporate the scaling laws at system-level and study the interaction between the scaled components, the energetic macroscopic representation formalism is employed. The novelty of the proposed method lies in structuring a scalable model and control for a reference electric axle to be used in system-level simulation. The novel organization consists of a refer-ence model and control complemented by two power adaptation elements at the electrical and mechanical sides. These latter elements consider the scaling effects, including the power losses. The methodology is applied for different case studies of battery electric vehicles, ranging from light to heavy-duty vehicles. Particular attention is paid to assessing the impact of the linear power-to-losses scaling method on energy consumption considering different power scaling factors and driving cycles, as compared to high-fidelity scaling methods.}},
  author       = {{Aroua, Ayoub}},
  keywords     = {{early development phases,efficiency,electric vehicles,energetic macroscopic representation,energetic study,experimental campaign,modularity,scalable modeling and control,scaling laws,system-level design}},
  language     = {{eng}},
  pages        = {{XXVIII, 246}},
  publisher    = {{Université de Lille ; Ghent University. Faculty of Engineering and Architecture}},
  school       = {{Ghent University}},
  title        = {{Scalability of electric axles for system-level design in the early development phases of electric vehicles}},
  year         = {{2023}},
}