Development of a Control System for a Series-Parallel Plug-In Hybrid Electric Vehicle

Abstract

This thesis outlines the development of a control system for a series-parallel plugin hybrid electric vehicle. The vehicle, developed at McMaster University for the EcoCAR 3 Advanced Vehicle Technology Competition, was produced in an effort to provide a Chevrolet Camaro with a high-performance, fuel efficient, hybrid powertrain. A rational design methodology was adopted and guided the development of the control system and the implementation of its respective algorithms. A simulation tool was created using MATLAB and Simulink which, in turn, allowed for the effectiveness of the supervisory control logic to be evaluated by approximating the vehicle’s energy consumption, fuel consumption, and emissions. The impact of hybridizing the vehicle’s powertrain was similarly assessed by comparing it against its unelectrified counterpart, the 2016 Chevrolet Camaro LT. A solution to the vehicle’s energy management problem was proposed in the form of an Adaptive Equivalent Consumption Minimization Strategy (A-ECMS) which was then evaluated against more common heuristic approaches as well as non-adaptive instantaneous minimization methods. An artificial neural network was selected as the strategy’s adaptation mechanism and it was used to identify specific vehicular driving patterns in real-time. The neural network addresses many issues that arise due to the sensitivity of algorithms that attempt to solve the energy management problem without prior knowledge of the driving cycle. The methods used during the process of the control system’s verification and calibration are also discussed in this thesis and, in addition, encompass the use of software representations of the vehicle’s Electronic Control Units (ECUs), the development of test cases, and the supervisory control software’s evaluation in the Model-in-the-Loop (MIL), Software-in-the-Loop (SIL), and Hardware-in-the-Loop (HIL) environments.