Session: 09-03-01 Current and Tidal Energy - 1

Paper Number: 81006

81006 - Control Co-Design of a Hydrokinetic Turbine With Open-Loop Optimal Control 

The design of a hydro-kinetic turbine (HKT) is often performed sequentially, with the physical design space (such as the blade geometry) being first explored with a focus on steady-state performance, followed by the control design addressing the transient responses once the physical parameters are defined. For highly dynamic and interactive systems such as HKT, the sequential design approach cannot fully explore the synergy between the physical entity and its control system dynamics, leading to sub-optimal design or overly constrained control solutions. Control co-design (CCD) provides a framework to simultaneously optimize physical design variables and control parameters for best performance and trade-offs.

 

The planned manuscript introduces a CCD framework for the rotor of an HKT. A coupled dynamic-hydrodynamic model is introduced, based on which the CCD problem is formulated to optimize the rotor power efficiency with time-variant flow profiles. The HKT CCD is investigated within the below-rated speed operating zone, and the objective is to maximize the HKT power production. Physical design parameters that are considered include twist angle and foil chord length along the blade. CCBlade, a blade element momentum theory based package, is used to model the hydrodynamics. Open-loop optimal torque control is applied for maximum power point tracking for the HKT. The trajectory is simulated and optimized using open-source software Dymos. 

 

Case studies for CCD with different control constraints are investigated together with sensitivity analysis to fluid profiles and initial geometries. Details, including physical designs, power production, trajectories of optimal control, and turbine rotating speed, are reported and discussed. Geometry and power performance of CCD design are compared against those of sequential design under the same constraints and flow profile. The results show that in the presence of control constraints, CCD leads to enhanced energy production, compared to sequential (design then control) methodology. Without control constraints, the CCD and the sequential design process will lead to identical physical design and power production results. 

 

The contributions of the work are four-fold: First, a CCD framework is developed for an HKT with open-loop optimal control. Second, with physical generator control load limit, the sequential design approach finds sub-optimal designs. Third, a comprehensive performance evaluation of the CCD results, compared with the sequential design, is carried out to reveal interactions between the two design spaces and provide insights for integration. Finally, the sensitivity of the CCD results to the flow profile and initial HKT geometry is analyzed. The computational footprint and numerical stability issues associated with the CCD implementation will also be discussed.

Presenting Author: Jing Sun University of Michigan

Authors:

Boxi Jiang University of Michigan
Mohammad Reza Amini University of Michigan
Yingqian Liao University of Michigan
Joaquim R. R. A. Martins University of Michigan
Jing Sun University of Michigan