Integrated optimisation for dynamic modelling, path planning and energy management in hybrid race vehicles
Lancaster University, Doctor of Philosophy, 2021
Supervisors: A. Montazeri and C.J. Taylor
Simulation software has for many years been developed to enhance the research and development phase of new vehicle introductions. With the introduction of the testing embargo in most forms of world championship motorsport, model validation is a necessity. To optimise the unknown vehicle and tyre parameters and to reduce the error between measured and simulated data in such a multi-input multi-output non-convex optimisation problem, a novel multi-objective particle swarm optimisation (PSO) technique is applied to ensure a fully validated vehicle model is developed and analysed for speed and performance.
These optimisation algorithms are further developed to explore the trajectory planning problem to improve the lap time for the shortest path, minimum curvature and a combined approach, producing optimal racing line pathways and vehicle dynamic inputs and output responses by exploring trajectories and vehicle traction circle limits.
Finally, a hybrid electric vehicle transient dynamics model for the control of energy management is presented. The hybrid powertrain contains an internal combustion engine, kinetic energy recovery system and heat energy recovery system with deployment and harvesting control parameters.
The performance of single-objective and multi-objective particle swarm optimisation algorithms are compared and analysed. The proposed simulation model and optimisation techniques are applied to address an array of problems, including model validation, racing line trajectory design, fastest lap time problem, and energy management strategies. All results are validated and optimised with respect to the experimental data collected on the real track in Silverstone to ensure the results can be applied to physical real-world scenarios.