CEVT uses cloud-based machine learning to improve life cycle for electric vehicle batteries in Big Battery Data project

Automotive brands have batteries on the streets worth billions. Managing them as an operating asset could mean billions saved. Find out how CEVT is using AI to improve the life cycle for electric vehicle batteries. 

In a project together with Chalmers University of Technology, CEVT has offered one of our battery engineers, Yizhou Zhang, to an Industrial PHD student position that will over a five-year period have as a target to use real-time data from cars on the roads and a cloud-based machine learning system to improve batteries for electric vehicles. Sven Sjöberg, Head of Battery Management and High Voltage Systems, tells us more. 

Tell us about this new PHD project?
The target is to use real-time data from cars on the roads and a cloud-based machine learning system for providing:

• Improved estimates of residual value of EV batteries
• Better life-time prediction
• Reduced warranty costs.
• Foresee and prevent malfunctions in EV-batteries.
• Offer smart services for planning of charging, climatization, maintenance etc.

Why are we interested in this as a company?
The battery constitutes half of the part cost of an electric vehicle today. We need to treat this investment as a working asset and investigate how we can make best use of it! This means optimizing part cost, quality costs and warranty costs and find ways to create additional cash flows – to monetize – from the energy and power system in the vehicle.

Who are involved?
Chalmers Department of Electronic Engineering and assistant professor Changfu Zou at the Automatic Control are responsible for the PhD project. CEVT employs and manages the Industrial PhD student, Yizhou Zhang. A cross-functional team from CEVTs departments of IT-infrastructure, AI, cloud-base applications, HV Battery and Electronics, Predictive Vehicle Control and Vehicle Integration, are supporting the project work and oversees the industrialization of the research results. 

We connect this project with PhD projects in related areas, for example at Nortical, a battery analytics startup collaborating with CEVT under the MobilityXlab - a collaboration hub with CEVT, Ericsson, Veoneer, Volvo Cars, Volvo Group & Zenseact.

What is the biggest challenge?
The biggest theoretical challenge is trying to predict lifetime for batteries that are completely new to the market and that have only been lifetime tested in a lab before. How will it behave in real-life operation? That is the question. 

The biggest practical challenge is to get enough field data with the right quality from real-world cars in the street. There are both technical, political, and legal roadblocks.

What is the greatest opportunity?
Geely brands have batteries on the streets worth many billions. Managing them as an operating asset rather than a liability could mean billions gained and saved for the Geely group.

What do we hope will come out of this once it is done? 
Apart from cost and profit improvements, long term this opens the door to optimize the design and usage of the energy storage and the power system – either for making smaller and cheaper batteries or to offer new electrically powered features to our customers.

What are you most proud of in this?
The team effort - that we together in the team saw an opportunity, grabbed it, recruited the right person, found the right partners, and got the government funding last year – and the strategic fit with the CEVT new mobility vision!

”I genuinely believe that electrification and digitalization are irresistible trends within the automotive industry. It will bring tremendous value to both the customer and society. I am delighted to have the opportunity to devote my five years of research time to such a meaningful topic,” says PhD student Yizhou Zhang. 

---