Energy management of hybrid and battery electric vehicles
File(s)
Author(s)
Nazemi, Arghavan
Type
Thesis
Abstract
This work focuses on improving the fuel economy of parallel Hybrid Electric Vehicles (HEVs) and dual-motor Electric Vehicles (EVs) through energy management strategies. Both vehicle models have two propulsion branches, each powering a separate axle: An engine and an electric motor in the HEV and two electric motors in the EV. This similarity in the vehicle models emphasises the need for similar energy management solutions. In Part Energy Management of HEVs of this thesis, a high-fidelity parallel Through-The-Road (TTR) HEV model is developed to study and test conventional control strategies. The traditional control strategies serve as a guide for developing novel heuristic control strategies. The Equivalent Consumption Minimisation Strategy (ECMS) is an optimisation-based control strategy used as the benchmark in this part of the work. A family of rule-based energy management strategies is proposed for parallel HEVs, including the Torque-levelling Threshold-changing Strategy (TTS) and its simplified version, the Simplified Torque-levelling Threshold-changing Strategy (STTS). The TTS applies a concept of torque-levelling, which ensures the engine works efficiently by operating with a constant torque as the load demand crosses a certain threshold, unlike the load-following approach commonly used. However, the TTS requires finely tuned constant torque and threshold parameters, making it unsuitable for real-time applications. To address this, two feedback-like updating laws are incorporated into the TTS to determine the constant torque and threshold online for real-time applications. Real-time versions of these strategies, Real-time Torque-levelling Threshold-changing Strategy (RTTS) and Real-time Simplified Torque-levelling Threshold-changing Strategy (RSTTS) are developed using a novel Driving Pattern Recognition (DPR) algorithm. The effectiveness of the RTTS is demonstrated by implementing it on a high-fidelity parallel hybrid passenger car and benchmarking it against ECMS. In Part Energy Management of EVs of the thesis, a low-fidelity model of a novel EV powertrain with two electric propulsion systems, one at each axle, has been developed to study and test its energy management with one of the main conventional optimal control methods, Dynamic Programming (DP). The EV model uses two differently sized traction motors at the front and rear axles. The thermal dynamics of the utilised Permanent Magnet Synchronous Motors (PMSMs) are studied. DP is first implemented onto the Baseline model that does not include any PMSM thermal dynamics, referred to as the Baseline DP, which acts as a benchmark since it is the conventional case. The thermal dynamics of the traction motors are then introduced in the second DP problem formulation, referred to as the Thermal DP, which is compared against the Baseline DP to evaluate the possible benefits of energy efficiency by the more informed energy management optimisation formulation. The best method is chosen to include these thermal dynamics in the overall energy management control strategy without significantly compromising computational time.
Version
Open Access
Date Issued
2022-10
Date Awarded
2023-06
Copyright Statement
Creative Commons Attribution NonCommercial Licence
License URL
Advisor
Evangelou, Simos
Publisher Department
Electrical and Electronic Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)