A Critical Examination of the Hysteresis in Wells Turbines Using Computational Fluid Dynamics and Lumped Parameter Models

Abstract

The hysteretic behavior of oscillating water column (OWC)-installed Wells turbines has been known for decades. The common explanation invokes the presence of unsteady aerodynamics due to the continuously varying incidence of the flow on the turbine blades. This phenomenon is neither new nor unique to Wells turbines, as an aerodynamic hysteresis is present in rapidly oscillating airfoils and wings, as well as in different types of turbomachinery, such as wind turbines and helicopter rotors, which share significant similarities with a Wells turbine. An important difference is the non-dimensional frequency: the hysteresis appears in oscillating airfoils only at frequencies orders of magnitude larger than the ones Wells turbines operate at. This work contains a re-examination of the phenomenon, using both computational fluid dynamics (CFD) and a lumped parameter model, and shows how the aerodynamic hysteresis in Wells turbines is negligible and how the often measured differences in performance between acceleration and deceleration are caused by the capacitive behavior of the OWC system. Results have been verified with respect to both spatial and temporal discretization, for unstalled and stalled operating conditions.

Influence of vortex interactions on the exergy transfer and performance of OWC wells turbine

“…In their study, the dynamic stall was responsible for the hysteresis effect under a high attack angle. The hysteresis under a low flow coefficient was due to the compressibility effects in the air chamber, which is consistent with the findings of Ghisu et al 17,18 It is generally accepted that the stall of Wells turbines is induced by the separation of the leading-edge (LE) boundary layer, rolling up to form leading-edge separation vortices (LEV) near the blade tip of the suction surface (SS), which is contrary to conventional turbines with the first separation near the hub owing to the small tangential velocity. There are two principal reasons for this phenomenon.…”

“…In the current study, the flow exergy loss caused by heat transport across the control surface is ignored. Meanwhile, the entropy generation _ S gen caused by viscous interactions was considered, as defined in equation (18), where μ mol and μ tur denote the molecular and turbulent viscosities, respectively, and Φ indicates the mechanical dissipation function.…”

Kinetics on the Integration of Methanol Aromatization with Raffinate Oil over ZSM-5/ZSM-11 Zeolite

“…Integral dynamics differential equations are solved by the fourth-order Runge–Kutta method. The simplex algorithm and the Levenberg–Marquardt optimization method are used to minimize the sum of squared residuals between the calculated value and the experimental value. − The objective function EOF is expressed as where X i exp , j is the experimental value of the dry basis mass fraction of component i under the experimental conditions of the j th group and X i calc , j is the calculated value of dry basis mass fraction of component i under the experimental conditions of group j . Through the MATLAB program, the kinetic experiment data are substituted into the kinetic ordinary differential equations for iterative calculation to obtain the reaction rate constant K i of the reaction at a reaction temperature of 360–430 °C.…”

“…As an intermediate product, olefins need to be independently set as a lumped. The target product of the reaction is aromatic hydrocarbons, so aromatic hydrocarbons are listed as separate lumps. − Finally, they divided the components in the MTA reaction system into five lumped. It is necessary to study the thermal and mass coupling reaction kinetics of methanol and light hydrocarbons, and the establishment of a lumped kinetic model can be a method to carry out related studies.…”

Discussion on “Influence of incoming wave conditions on the hysteretic behavior of an oscillating water column system for wave energy conversion” by J. Peng, C. Hu and C. Yang

“…In their paper [1], Peng, Hu and Yang presented a lumped parameter model (LPM) to quantify the hysteresis in Oscillating Water Column (OWC) systems. We could not avoid noticing that the assumptions at its origin, the derivation of the LPM and the methodology used to obtain an analytical solution (after linearization) are remarkably similar to the ones presented in some of our published works [2,3,4], none of which were referenced in [1]. The similarities between the model presented by Peng and his coauthors and the model published by ourselves are significant, to the point that we find hard to believe that they could have derived it authonomously.…”