Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Dynamic Performance Improvement of Non-Isolated DC-DC converters and PV Energy Harvesting Systems using One Step Finite Control Set Model Predictive Control coupled with geometrical domain analysis

Harzig, Thibaut (2022) Dynamic Performance Improvement of Non-Isolated DC-DC converters and PV Energy Harvesting Systems using One Step Finite Control Set Model Predictive Control coupled with geometrical domain analysis. Doctoral Dissertation, University of Pittsburgh. (Unpublished)

Download (5MB) | Preview


The proposed research project uses the well-documented Model Predictive Control (MPC) framework to improve the dynamic performance of power converters and differential power processing (DPP) architecture with the help of geometrical domain analysis. The first research task is to implement a One Step Finite Control Set MPC (FCS-MPC) for a DC-DC boost converter. In the proposed control scheme, the cost function is built using time-optimal trajectories of the boost converter to solve the convergence issue brought by the non-minimum phase behavior of this system. The constraint of the proposed FCS-MPC limits important voltage deviations when using time-optimal trajectories. The second research task aims at proposing a generalized One Step FCS-MPC for the most common Non-Isolated DC-DC converters (buck boost and buck-boost). The proposed control scheme uses a unified switching model of non-isolated DC-DC converters and adapts the existing time-optimal boundary controllers to the FCS-MPC framework. The contributions are the avoidance of non-minimum phase issues, the limitation of voltage deviation and current spikes, and the possibility to target a specific steady state switching frequency. The third research task involves the implementation of FCS-MPC control schemes for the Differential Power Processing (DPP) PV-bus direct architecture. In this architecture a bidirectional Flyback is connected in parallel with each PV panel to operate a maximum power point tracking (MPPT). This system incorporates a string converter controlling the PV string current minimizing the power processed by bidirectional Flyback converters with a Least Power Point Tracking (LPPT) algorithm. In this research task the classical direct duty cycle control MPPT is replaced by a FCS-MPC MPPT using geometrical domain analysis. The classical LPPT implemented with PI controllers is replaced by a FCS-MPC where the cost function is the power processed by the bidirectional flyback converters. The benefit is to avoid interactions between LPPT and MPPTs with an increase in the control dynamic performance. Overall, the proposed set of control schemes improves the minimization of power stress on bidirectional Flyback converters.


Social Networking:
Share |


Item Type: University of Pittsburgh ETD
Status: Unpublished
CreatorsEmailPitt UsernameORCID
Harzig, ThibautTHH39@pitt.eduTHH39
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Committee ChairGrainger, Brandonbmg10@pitt.edubmg10
Committee MemberBajaj, NikhilNBAJAJ@pitt.eduNBAJAJ
Committee MemberKerestes, Robertrjk39@pitt.edurjk39
Committee MemberMao, Zhi-Hongzhm4@pitt.eduzhm4
Committee MemberKwasinski, Alexisakwasins@pitt.eduakwasins
Date: 6 September 2022
Date Type: Publication
Defense Date: 21 June 2022
Approval Date: 6 September 2022
Submission Date: 1 August 2022
Access Restriction: No restriction; Release the ETD for access worldwide immediately.
Number of Pages: 160
Institution: University of Pittsburgh
Schools and Programs: Swanson School of Engineering > Electrical and Computer Engineering
Degree: PhD - Doctor of Philosophy
Thesis Type: Doctoral Dissertation
Refereed: Yes
Uncontrolled Keywords: Model Predictive Control, Time optimal Control, Non-Isolated DC-DC converters, Buck, Boost, Buck-Boost, Power Stress Minimization, Differential power processing architecture, PV elements
Date Deposited: 06 Sep 2022 16:41
Last Modified: 06 Sep 2022 16:41


Monthly Views for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item