Advanced Optimization Methods for The Design of Aluminium Based Battery Enclosures for Electric Vehicles
Lead Research Organisation:
University of Oxford
Department Name: Engineering Science
Abstract
Description
Climate change is one of the most important topics in the 21st century. The transportation sector, a major emitter of CO2, must adapt to reduce its impact on the environment. As a part of this sector, the automotive industry is focusing on electric vehicles as one of its strategies to reduce CO2 emissions. As the electric vehicle market grows, manufacturers are developing new vehicle platforms specifically for electric vehicles.
One of the challenges is compensating for the added weight of batteries by decreasing vehicle body weight. This is especially important because the additional battery mass reduces the range of an electric vehicle. Thus, lightweight engineering is a crucial part of the vehicle's design, including the battery enclosure's design. Aluminium extrusion profiles for battery enclosures are one of the preferred structures subjected to crash loads, due to their high energy absorption to weight ratio.
The design process of such battery enclosures is conducted and expedited using advanced computer-aided engineering tools, cutting down physical tests. Yet, the process still heavily relies on the designer's expertise. Advanced computer-aided engineering tools targeting optimised solutions, based on finer design details like material choice, can accelerate market entry. This, enabling designers to effectively design and optimize battery enclosures, considering material properties and various assembly methods such as welding, bonding, or mechanical fastening. Therefore, enabling lightweight structures that save material and resources. For this purpose, an adopted methodology must be developed.
Aims and objectives
The proposed research aims to develop an optimised design process for electric vehicle battery enclosures. The research is interdisciplinary bridging several disciplines, namely materials science, engineering science, and computer science.
Focusing on critical aspects of materials engineering, impact and shock performance of material systems, and optimal design to enhance the topology of aluminium extrusion profile.
Applying machine learning techniques for speed improvements of the structure's topology optimisation under consideration of the material properties and impact specific characteristics such as strain rate dependency.
The proposed investigation will rely upon stochastic approaches to represent the mechanical performance of materials systems at multiple length scales. The proposed objectives help to enhance and accelerate the design process of battery enclosure structures.
Novelty of the research methodology
Physics-informed neural networks, capable of capturing the underlying mechanics of the problem, are often considered more advanced than purely data-driven methods. By incorporating properties like the material system, rate-dependent material behaviour, and other properties, these networks can notably speed up the topology optimization process while considering the physics. This enables faster optimisation and design of structures subjected to impact loads.
Additionally, by conducting a variability analysis, we can gauge the impact of multiple structural properties, including the manufacturing process and rate-dependent material behaviour, on the structure's topology.
The proposed methodology aims to expedite and improve the design of battery enclosure profiles while ensuring they meet crash safety requirements.
EPSRC
This project falls within the EPSRC engineering design research area.
Involved partner
The EPSRC iCase involves the industrial partner Constellium.
Climate change is one of the most important topics in the 21st century. The transportation sector, a major emitter of CO2, must adapt to reduce its impact on the environment. As a part of this sector, the automotive industry is focusing on electric vehicles as one of its strategies to reduce CO2 emissions. As the electric vehicle market grows, manufacturers are developing new vehicle platforms specifically for electric vehicles.
One of the challenges is compensating for the added weight of batteries by decreasing vehicle body weight. This is especially important because the additional battery mass reduces the range of an electric vehicle. Thus, lightweight engineering is a crucial part of the vehicle's design, including the battery enclosure's design. Aluminium extrusion profiles for battery enclosures are one of the preferred structures subjected to crash loads, due to their high energy absorption to weight ratio.
The design process of such battery enclosures is conducted and expedited using advanced computer-aided engineering tools, cutting down physical tests. Yet, the process still heavily relies on the designer's expertise. Advanced computer-aided engineering tools targeting optimised solutions, based on finer design details like material choice, can accelerate market entry. This, enabling designers to effectively design and optimize battery enclosures, considering material properties and various assembly methods such as welding, bonding, or mechanical fastening. Therefore, enabling lightweight structures that save material and resources. For this purpose, an adopted methodology must be developed.
Aims and objectives
The proposed research aims to develop an optimised design process for electric vehicle battery enclosures. The research is interdisciplinary bridging several disciplines, namely materials science, engineering science, and computer science.
Focusing on critical aspects of materials engineering, impact and shock performance of material systems, and optimal design to enhance the topology of aluminium extrusion profile.
Applying machine learning techniques for speed improvements of the structure's topology optimisation under consideration of the material properties and impact specific characteristics such as strain rate dependency.
The proposed investigation will rely upon stochastic approaches to represent the mechanical performance of materials systems at multiple length scales. The proposed objectives help to enhance and accelerate the design process of battery enclosure structures.
Novelty of the research methodology
Physics-informed neural networks, capable of capturing the underlying mechanics of the problem, are often considered more advanced than purely data-driven methods. By incorporating properties like the material system, rate-dependent material behaviour, and other properties, these networks can notably speed up the topology optimization process while considering the physics. This enables faster optimisation and design of structures subjected to impact loads.
Additionally, by conducting a variability analysis, we can gauge the impact of multiple structural properties, including the manufacturing process and rate-dependent material behaviour, on the structure's topology.
The proposed methodology aims to expedite and improve the design of battery enclosure profiles while ensuring they meet crash safety requirements.
EPSRC
This project falls within the EPSRC engineering design research area.
Involved partner
The EPSRC iCase involves the industrial partner Constellium.
People |
ORCID iD |
| Nikica Petrinic (Primary Supervisor) | |
| Jan Wittig (Student) |
Studentship Projects
| Project Reference | Relationship | Related To | Start | End | Student Name |
|---|---|---|---|---|---|
| EP/Y52878X/1 | 01/10/2023 | 30/09/2028 | |||
| 2891974 | Studentship | EP/Y52878X/1 | 01/10/2023 | 30/09/2027 | Jan Wittig |