Postdoc: ML-Driven Solid-State Battery Estimation & Control

Kungliga Tekniska högskolan; Stockholms kommun; On-site; SEK 320,000 - 420,000; Full time

Job summary

KTH Royal Institute of Technology invites applications for a two-year postdoctoral researcher position in solid-state batteries under the COMPEL program. The project blends physics-based modelling with machine learning for state estimation and predictive control, aiming to understand how charging conditions, temperature and mechanical load influence degradation and lifetime.

The role involves developing reduced-order models, integrating ML with physics, and collaborating across related COMPEL

Qualifications

• Proficiency in written and spoken English.
• Strong theoretical background in systems modelling, estimation, optimization or machine learning.
• Awareness of diversity and equal treatment issues with focus on gender equality.

Responsibilities

• Develop reduced-order electro-chemo-mechanical models and state estimation methods for solid-state batteries.
• Integrate physics-based modelling with machine learning for parameter estimation and degradation prediction.
• Analyze electrochemical and structural characterization data, including X-ray CT image reconstruction and analysis.
• Explore how charging protocols, temperature and mechanical loading affect performance and lifetime.
• Collaborate with parallel COMPEL projects and contribute to health-aware battery management.

Skills

English proficiency

Systems modelling

Estimation

Optimization

Machine learning

Education

Doctoral degree in Control Engineering or related field

Job description

KTH Royal Institute of Technology invites applications for a two-year postdoctoral researcher position in solid-state batteries under the COMPEL program. The project blends physics-based modelling with machine learning for state estimation and predictive control, aiming to understand how charging conditions, temperature and mechanical load influence degradation and lifetime.

The role involves developing reduced-order models, integrating ML with physics, and collaborating across related COMPEL