In this research, control co-design was used to study the impact of the inclusion of key vehicle performance criteria on the optimal design and control of PHEV powertrains. In particular, first, a moderate-fidelity PHEV powertrain system was modeled and then optimized using both a performance-based and a non-performance-based control co-design formulation. Thereafter, the final results from the two formulations were compared. Based on the results, it was revealed that the inclusion of such criteria profoundly changes the design and supervisory control of the final system such that ignoring these criteria in the design stage could lead to a final suboptimal system.
The most common formulation of control co-design, the all-in-once, or the AIO formulation, though effective for small systems, can become impractical to apply to large interconnected systems with two or more subsystems. This is generally due to the complexity of the optimization problems for such systems and/or due to a desired hierarchical design paradigms for such systems. To address this, a novel decomposition-based formulation of control co-design was proposed in this work and its practicality was assessed via a control co-design problem for a PHEV powertrain. The results revealed that the new formulation was capable of finding the system-level optimal solutions, and that it can play a critical role in the design of such systems when the AIO formulations are not practical to apply.
Design and parametric uncertainties within dynamic systems if left ignored, can cause system performance degradation or even failure in real-world conditions. These uncertainties are addressed via reliability-based CCD (RBCCD) formulations. Though, since current RBCCD formulations present some limitations in their accuracy, their applications are limited. In this research, a new set of RBCCD formulations that address these limitations were proposed. In addition, different reliability analysis methods , such as generalized polynomial chaos- and MPP-based algorithms for the new RBCCD formulations were developed and compared.