Energy Optimization using Model Predictive Control for Engine Cooling Systems in Heavy Duty Trucks
Open Access
- Author:
- Huynh, Brian
- Graduate Program:
- Mechanical Engineering
- Degree:
- Master of Science
- Document Type:
- Master Thesis
- Date of Defense:
- May 29, 2019
- Committee Members:
- Stephanie Stockar, Thesis Advisor/Co-Advisor
Christopher D. Rahn, Committee Member - Keywords:
- Model Predictive Control
Engine Cooling System
Energy Optimization - Abstract:
- To comply with stringent emission and fuel consumption standards, manufacturers in the automotive industry have strived to improve the fuel economy of their vehicles through the advancements in the powertrain, drivetrain, and aerodynamics. Improvements in the fuel economy can also be obtained from reducing the auxiliary load exerted on the engine from components such as the alternator, water pumps, and radiator fan. The engine cooling system contributes to the auxiliary loads by utilizing components, such as the water pump and radiator fan, to control the temperature of the engine. This Thesis investigates the fuel economy improvements in the engine cooling system for heavy-duty trucks by designing a predictive control strategy for the radiator fan. First, a model of the engine cooling system is derived and validated through experimental tests. Then, an optimal control problem is formulated with the objective of optimizing two performance metrics: 1) vehicle fuel consumption including auxiliaries and 2) the coolant temperature tracking performance. Due to the nonlinearities in the system, Dynamic Programming (DP) is used as the solution algorithm. Once the analysis of the DP solution is performed and a benchmark is governed, a real-time implementable strategy is developed using Model Predictive Control (MPC). The advantage of MPC, namely optimality while satisfying constraints, can be exploited further in the framework of Connected and Automated Vehicles (CAVs), where information on future driving conditions become available. In this Thesis, the optimal fan control for a connected truck is developed and tested in multiple environments, such as Software-in-the-Loop and Engine-in-the-Loop. The proposed predictive control strategy has obtained average fuel savings of 1.84\% through Engine-in-the-Loop testing.