Optimal control applied to thermal management

Optimal predictive control applied to thermal management of electrified heavy vehicles


Thesis Background


At Volvo, we are committed to the ambitions and climate change goals of the Paris Agreement. Environmental aspects, customer demand and the rapid development in transport industry have created a competitive environment in providing better transportation solutions. The need for energy efficiency and improving the lifetime of the component, specifically in electrified heavy-duty trucks, drives the industry in utilizing advanced optimization-based approaches. To this day, the main bottleneck for bringing trust to industry for implementing optimization-based methods, is the computational complexity and memory footprint. To overcome this challenge, there has been substantial research on developing embeddable solution methods for optimal control problems. Even though deployment of the commercial solutions for thermal management can be a time-consuming task, defining a suitable and feasible problem is the ambition of this work. This thesis is hoped to be performed in close collaboration with the other thesis work that focuses on the model development of the cooling system for optimal control problem.


Problem motivating the project


The thermal management problem for electrified vehicles like most problems is a compromise between performance and energy consumption. That means, on one side it is desirable to utilize the cooling power as less as possible, and on the other side, it is to keep the component temperature as low as possible. To make the matter even more challenging, one should consider the fact that both the cooling power and the temperature of the components are bounded. Violation of these bounds sometimes is not physically possible (e.g., the cooling power is limited to a fixed value) and sometimes is not desirable as the violation can damage the component: temperature of the battery packs should stay within a specific range and going beyond that will decrease the lifetime.
The main challenge here is to define the optimal control problem in a way that all the requirements from different components are addressed. Some of these requirements are straightforward, as the temperature bounds or the limit on the active cooling power, while some of the requirements are more complicated to bring into the optimal control problem, e.g., the fan power consumption, or the aging model for the batteries. Some of these terms are nonlinear and it is important to formulate them in a way that the feasibility of the problem is kept.


Objective or Research Question


How to control the complete cooling system and thermal management for, e.g., motor drive system, battery packs, active/passive cooling component, pumps, and valves, in Simulink in a receding horizon control strategy using a quadradic program?
How to address different cost terms in a balanced manner? For example, is it possible to convert all these terms in the same unit so that the tuning work is simplified?
Some valves are on-off, and without over-complicating the problem (e.g., going into event-based optimal theory), is it possible to control the valves as continuous control signal?


Deliverables (flexible)
Formulation of the optimization problem
Analysis of the aging model for the batteries
Mathematical models with time-varying parameters of heavy vehicle cooling systems. The types of vehicles include battery-electric and fuel-cell-electric.
Robustness analysis of the estimation results for being utilized in a receding horizon control strategy.
System identification of the components heat loss model including the dependency on power intake and coolant temperature.
Design a real-time optimal control strategy for fallback mode

Requirement on student background:
Master students in Automotive, Control, Mechatronics or Engineering Physics.
Knowledge of system identification, MATLAB/Simulink. It is a bonus to have some understanding of fluid dynamics.
Please submit your CV and motivation letter.

Supervision and examination:
Volvo Group, Powertrain Strategic Development.
Chalmers University of Technology
Thesis Level: Master
Language: English
Starting date: February 2023
Number of students: 1 or 2
Physical location:
Students are welcome to sit at Volvo Lundby.


Kindly note that due to GDPR, we will not accept applications via mail. Please use our career site.