Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


36819.pdf (2.52 MB)

ETD Abstract Container

Abstract Header

Combined Design and Control Optimization of Stochastic Dynamic Systems

Azad, Saeed
http://orcid.org/0000-0002-4917-6336
http://rave.ohiolink.edu/etdc/view?acc_num=ucin1602153122063302

Abstract Details

2020, PhD, University of Cincinnati, Engineering and Applied Science: Mechanical Engineering.
Optimization of dynamic engineering systems requires an integrated approach that accounts for the coupling between embodiment design and control system design, simultaneously. Generally known as combined design and control optimization (co-design), these methods offer superior system performance and reduced costs. Despite the widespread use of co-design techniques in the literature, extremely limited research has been done to address the issue of uncertainty in co-design problem formulations. This is problematic as all engineering models contain some level of uncertainty that might negatively affect the system’s performance, if overlooked. Accounting for these uncertainties transforms the deterministic problem into a stochastic one, requiring the use of appropriate stochastic optimization approaches. Therefore, this dissertation serves as the starting point for research on stochastic co-design problems when the uncertainty is propagated into the system from random design decision variables and/or problem parameters. Specifically, a simultaneous co-design formulation within multidisciplinary dynamic system design optimization (MDSDO), along with a special class of direct methods, known as direct transcription (DT), are consistently used throughout this research as the basis for uncertainty considerations. Using techniques from robust design optimization (RDO), we develop a novel stochastic co-design formulation within MDSDO, known as robust MDSDO (R-MDSDO). This formulation enables a protective measure against uncertainties by minimizing the sensitivity of the objective function to variations in design decision variables and fixed problem parameters. The R-MDSDO formulation is applied to two case studies to assess its effectiveness and implementation challenges. A more rigorous evaluation of probabilistic constraints is required to ensure reliability. In this dissertation, we develop a novel stochastic co-design approach based on the principles of RBDO. We implement the reliability analysis through a performance measure approach (PMA) and employ a double-loop most-probable-point (MPP) method, along with a first-order reliability method (FORM) to evaluate the probabilistic constraints. This novel formulation is known as the double-loop reliability-based MDSDO (RB-MDSDO). A major concern in the implementation of the proposed double-loop RB-MDSDO is its high computational cost for the computationally-intensive co-design problems. Therefore, we use the sequential optimization and reliability assessment (SORA) algorithm to develop a novel single-loop MPP method for RB-MDSDO. In this algorithm, the optimization and reliability assessment steps are decoupled from each other and run sequentially. The effectiveness, efficiency, and scalability of the proposed RB-MDSDO formulations are assessed by solving the complex co-design problem of an automotive active-suspension system. By comparing the solutions of the double-loop and the single-loop RB-MDSDO, we concluded that the latter approach offers superior computational efficiency for stochastic co-design problems while preserving accuracy. Finally, the proposed formulations are used to address the real-world complex co-design problem of hybrid-electric vehicle (HEV) powertrain under uncertainty. In this dissertation, we use R-MDSDO, as well as the RB-MDSDO to explicitly account for random variations within design decision variables and fixed problem parameters for a power-split HEV powertrain. The impact of these uncertainties within the HEV powertrain model and problem formulation is then demonstrated by comparing the results from R-MDSDO and RB-MDSDO to those from the associated deterministic co-design problems.
Michael Alexander-Ramos, Ph.D. (Committee Chair)
Kelly Cohen, Ph.D. (Committee Member)
Manish Kumar, Ph.D. (Committee Member)
David Thompson, Ph.D. (Committee Member)
125 p.
Mechanical Engineering
Stochastic dynamic systems; combined design and control optimization; MDSDO; Robust design optimization; Reliability-based design optimization; HEV Powertrain

Recommended Citations

Citations

  • Azad, S. (2020). Combined Design and Control Optimization of Stochastic Dynamic Systems [Doctoral dissertation, University of Cincinnati]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1602153122063302

    APA Style (7th edition)

  • Azad, Saeed. Combined Design and Control Optimization of Stochastic Dynamic Systems. 2020. University of Cincinnati, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=ucin1602153122063302.

    MLA Style (8th edition)

  • Azad, Saeed. "Combined Design and Control Optimization of Stochastic Dynamic Systems." Doctoral dissertation, University of Cincinnati, 2020. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1602153122063302

    Chicago Manual of Style (17th edition)

ucin1602153122063302
179
© 2020, all rights reserved.
This open access ETD is published by University of Cincinnati and OhioLINK.