ProjectConvex optimization

Application Areas
Competence Areas at RISE ICT
Convex optimization

A powerful tool that helps Volvo AB to understand the future of transportation

Imagine yourself that you are in charge of the research and development budget for Volvo AB. What types of vehicles should you develop for the year 2030 and how will you make money during the expected transition towards increasingly energy efficient powertrains?  Could the introduction of hybrid electric vehicles be the first step in a gradual development towards increasingly more electrified powertrains or will internal combustion engine still be the most cost efficient solution?

To help the vehicle industry to make these strategic decisions there is at the department of Control Automation and Mechatronics an ongoing development of methods and software tools for system studies of hybrid vehicles. Of particular interest is the control strategy that decides the power flow in and out of the battery. It is well known that the tuning and structure of this power flow control strategy strongly influences the energy efficiency of the hybrid vehicle.

Spurred by Professor Bo Egardt’s course in Convex optimization a breakthrough was achieved in 2010 by Nikolce Murgovski and Lars Johannesson Mårdh making it possible to simultaneously size the energy buffer and to find the optimal power flow control strategy in a convex optimization problem. Compared to previous methods based on Dynamic Programming the convex optimization approach makes it possible to include models with more dynamics, opening up possibilities of including battery wear models in the optimization of the power flow control strategy. The convex optimization approach is now developed further in Nikolce’s post doc with the research aim to cover most types of powertrain technologies and hybrids with supercapacitors, pneumatic or hydraulic compression storages and mechanical flywheels.

Although the convex optimization approach for hybrid vehicle sizing and control is still at an early stage the developed software and methods were used during the fall of 2012 in a project at Viktoria Swedish ICT and Volvo AB. The aim of the project was to study which types of powertrain technologies will be the most cost efficient for Volvo’s most important vehicle types and markets. The study was based on numerous scenarios for the price development for fuel, electricity and powertrain component prices.

Jonas Hellgen, the project leader at Volvo says: "The convex optimization approach is computationally efficient, hence it is possible to investigate far more powertrain variants and price scenarios that would otherwise be possible. The results from this study and future similar studies may influence the chosen future powertrain technology of the Volvo group".

The picture above shows the project team in the study, from the left: Esteban Gelso (Volvo AB), Jonas Hellgren (Volvo AB), Lars Johannesson Mårdh (Viktoria Swedish ICT) and Nikolce Murgovski (Viktoria Swedish ICT and Chalmers).