Showing papers in "Physics of Fluids in 2019"

Machine learning methods for turbulence modeling in subsonic flows around airfoils

Abstract: In recent years, the data-driven turbulence model has attracted widespread concern in fluid mechanics. The existing approaches modify or supplement the original turbulence model by machine learning based on the experimental/numerical data, in order to augment the capability of the present turbulence models. Different from the previous researches, this paper directly reconstructs a mapping function between the turbulent eddy viscosity and the mean flow variables by neural networks and completely replaces the original partial differential equation model. On the other hand, compared with the machine learning models for the low Reynolds (Re) number flows based on direct numerical simulation data, high Reynolds number flows around airfoils present the apparent scaling effects and strong anisotropy, which induce large challenges in accuracy and generalization capability for the machine learning algorithm. We mainly concentrate on the high Reynolds number turbulent flows around the airfoils and take the results calculated by the computational fluid dynamics solver with the Spallart-Allmaras (SA) model as training data to construct a high-dimensional data-driven network model based on machine learning. The radial basis function neural network and the auxiliary optimization methods are adopted, and the individual models are built separately for the flow fields of the near-wall region, wake region, and far-field region. The training data in this paper is extracted from only three subsonic flow fields of NACA0012 airfoil. The data-driven turbulence model can be applied to various airfoils and flow states, and the predicted eddy viscosity, lift/drag coefficients, and skin friction distributions are all in good agreement with the results of the original SA model. This research demonstrates the promising prospect of machine learning methods in future studies about turbulence modeling.

Fast flow field prediction over airfoils using deep learning approach

Abstract: In this paper, a data driven approach is presented for the prediction of incompressible laminar steady flow field over airfoils based on the combination of deep Convolutional Neural Network (CNN) and deep Multilayer Perceptron (MLP). The flow field over an airfoil depends on the airfoil geometry, Reynolds number, and angle of attack. In conventional approaches, Navier-Stokes (NS) equations are solved on a computational mesh with corresponding boundary conditions to obtain the flow solutions, which is a time consuming task. In the present approach, the flow field over an airfoil is approximated as a function of airfoil geometry, Reynolds number, and angle of attack using deep neural networks without solving the NS equations. The present approach consists of two steps. First, CNN is employed to extract the geometrical parameters from airfoil shapes. Then, the extracted geometrical parameters along with Reynolds number and angle of attack are fed as input to the MLP network to obtain an approximate model to predict the flow field. The required database for the network training is generated using the OpenFOAM solver by solving NS equations. Once the training is done, the flow field around an airfoil can be obtained in seconds. From the prediction results, it is evident that the approach is efficient and accurate.

Direct simulation Monte Carlo on petaflop supercomputers and beyond

Abstract: The gold-standard definition of the Direct Simulation Monte Carlo (DSMC) method is given in the 1994 book by Bird [Molecular Gas Dynamics and the Direct Simulation of Gas Flows (Clarendon Press, Oxford, UK, 1994)], which refined his pioneering earlier papers in which he first formulated the method. In the intervening 25 years, DSMC has become the method of choice for modeling rarefied gas dynamics in a variety of scenarios. The chief barrier to applying DSMC to more dense or even continuum flows is its computational expense compared to continuum computational fluid dynamics methods. The dramatic (nearly billion-fold) increase in speed of the largest supercomputers over the last 30 years has thus been a key enabling factor in using DSMC to model a richer variety of flows, due to the method’s inherent parallelism. We have developed the open-source SPARTA DSMC code with the goal of running DSMC efficiently on the largest machines, both current and future. It is largely an implementation of Bird’s 1994 formulation. Here, we describe algorithms used in SPARTA to enable DSMC to operate in parallel at the scale of many billions of particles or grid cells, or with billions of surface elements. We give a few examples of the kinds of fundamental physics questions and engineering applications that DSMC can address at these scales.

A deep learning enabler for nonintrusive reduced order modeling of fluid flows

Abstract: In this paper, we introduce a modular deep neural network (DNN) framework for data-driven reduced order modeling of dynamical systems relevant to fluid flows. We propose various DNN architectures which numerically predict evolution of dynamical systems by learning from either using discrete state or slope information of the system. Our approach has been demonstrated using both residual formula and backward difference scheme formulas. However, it can be easily generalized into many different numerical schemes as well. We give a demonstration of our framework for three examples: (i) Kraichnan-Orszag system, an illustrative coupled nonlinear ordinary differential equation, (ii) Lorenz system exhibiting chaotic behavior, and (iii) a nonintrusive model order reduction framework for the two-dimensional Boussinesq equations with a differentially heated cavity flow setup at various Rayleigh numbers. Using only snapshots of state variables at discrete time instances, our data-driven approach can be considered truly nonintrusive since any prior information about the underlying governing equations is not required for generating the reduced order model. Our a posteriori analysis shows that the proposed data-driven approach is remarkably accurate and can be used as a robust predictive tool for nonintrusive model order reduction of complex fluid flows.In this paper, we introduce a modular deep neural network (DNN) framework for data-driven reduced order modeling of dynamical systems relevant to fluid flows. We propose various DNN architectures which numerically predict evolution of dynamical systems by learning from either using discrete state or slope information of the system. Our approach has been demonstrated using both residual formula and backward difference scheme formulas. However, it can be easily generalized into many different numerical schemes as well. We give a demonstration of our framework for three examples: (i) Kraichnan-Orszag system, an illustrative coupled nonlinear ordinary differential equation, (ii) Lorenz system exhibiting chaotic behavior, and (iii) a nonintrusive model order reduction framework for the two-dimensional Boussinesq equations with a differentially heated cavity flow setup at various Rayleigh numbers. Using only snapshots of state variables at discrete time instances, our data-driven approach can be considered trul...

Experimental study of the parameters for stable drop-on-demand inkjet performance

Abstract: We present an experimental study of drop-on-demand inkjet behavior, with particular emphasis on the thresholds for drop generation and formation of satellite drops, using inks covering a range of fluid properties. Drop behavior can be represented as a “phase diagram” in a parameter space bound by the dimensionless number Z (the inverse of the Ohnesorge number) and the Weber number of the fluid jet prior to drop formation, Wej. Stable drop generation is found to be bounded by a parallelogram with minimum and maximum values of 2 4, the value of Wej at which satellite drops form decreases with increasing Z until at very large values of Z single drops can no longer form at any Wej. However, despite the large range of fluid properties over which stable drops can form, the need for a large range of both Z and Wej limits the region of practical ink design to the approximate range of 2 < Z < 20. These results are shown to be compatible with current models of the drop formation process reported in the literature.

New normalized Rortex/vortex identification method

Abstract: A new vortex identification criterion, named ΩR, is proposed for the normalization of Rortex, using the idea of the Omega method (Ω). ΩR is a normalized function from 0 to 1, which measures the relative rotation strength on the plane perpendicular to the local rotation axis. The advantages of the proposed ΩR method can be summarized as follows: (1) ΩR is from 0 to 1 and can be further used in statistics and correlation analysis as a physical quantity. (2) ΩR can distinguish the rotational vortices from the shear layers, discontinuity structures, and other non-physical structures. (3) ΩR is quite robust and can be always set as 0.52 to capture vortex structures in different cases and at different time steps.

Accelerating deep reinforcement learning strategies of flow control through a multi-environment approach

Abstract: Deep Reinforcement Learning (DRL) has recently been proposed as a methodology to discover complex active flow control strategies [Rabault et al., “Artificial neural networks trained through deep reinforcement learning discover control strategies for active flow control,” J. Fluid Mech. 865, 281–302 (2019)]. However, while promising results were obtained on a simple 2-dimensional benchmark flow at a moderate Reynolds number, considerable speedups will be required to investigate more challenging flow configurations. In the case of DRL trained with Computational Fluid Dynamics (CFD) data, it was found that the CFD part, rather than training the artificial neural network, was the limiting factor for speed of execution. Therefore, speedups should be obtained through a combination of two approaches. The first one, which is well documented in the literature, is to parallelize the numerical simulation itself. The second one is to adapt the DRL algorithm for parallelization. Here, a simple strategy is to use several independent simulations running in parallel to collect experiences faster. In the present work, we discuss this solution for parallelization. We illustrate that perfect speedups can be obtained up to the batch size of the DRL agent, and slightly suboptimal scaling still takes place for an even larger number of simulations. This is, therefore, an important step toward enabling the study of more sophisticated fluid mechanics problems through DRL.

Mechanism study of coupled aerodynamic and thermal effects using plasma actuation for anti-icing

Abstract: Anti-icing performance using the surface dielectric barrier discharge plasma actuator is studied using detailed visualization and surface thermal measurements. To reveal the physical mechanism of coupled aerodynamic and thermal effects on anti-icing, three types of actuators are designed and mounted on a NACA 0012 airfoil. The coupled aerodynamic and thermal effects are confirmed in still air. The results show that the plasma actuation is effective for in-flight anti-icing, and the anti-icing performance is directly related to the design of the plasma actuators based on the coupled aerodynamic and thermal effects. When the direction of plasma induced flow is consistent with the incoming flow, the heat generated by plasma discharge is concentrated in the region of the actuator and the ability of the actuator for heat transfer downstream is relatively weak during the anti-icing. When the induced flow is opposite to the incoming flow, there is less heat accumulation in the actuator region, while the ability of heat transfer downstream becomes stronger. With the consistent and opposite direction of induced flow, the plasma actuation can ensure that 57% and 81% chord of the lower surface of the airfoil are free of the ice accumulation, respectively. Another actuator is designed to induce the air jets approximately perpendicular to the airfoil surface. This exhibits both a stronger ability of heat accumulation locally and heat transfer downstream and hence ensures that there is no ice on the entire lower surface of the airfoil.

Natural convection of water-based carbon nanotubes in a partially heated rectangular fin-shaped cavity with an inner cylindrical obstacle

Abstract: This study investigates the heat augmentation and hydromagnetic flow of water-based carbon nanotubes (CNTs) inside a partially heated rectangular fin-shaped cavity. A thin heated rod is placed within the cavity to create a resistance or to provide a source for heat transfer. The obstacle is tested for the heated case, while the right side of the horizontal tip is tested for three different temperatures (adiabatic, cold, and heated). The left vertical side of the cavity is partially heated with temperature Th, and the rest of the sides are kept cold at temperature Tc except the right tip. Two different thermal boundary conditions (prescribed temperature and adiabatic) are employed on the fin tip. The CNTs and water are assumed to be in thermal equilibrium with no-slip velocity. The magnetic field and thermal radiation are introduced in the momentum and energy equations, respectively. The governing equations are obtained in dimensionless form by means of dimensionless variables. The numerical computation is performed via the finite element method using the Galerkin approach. The substantial effects of emerging parameters on the streamlines, isotherms, dimensionless velocities, and temperature are reported graphically and discussed. In the case of a cold or adiabatic fin-tip, a drop to minimum is found in the dimensionless temperature. The components of velocity are perceived maximum at a vertical corner while minimum at the horizontal corner. It is demonstrated that the local Nusselt numbers are increased by introducing both solid volume fraction of CNTs and radiation effects, while the Nusselt number noticed maximum at the corners.

Modified normalized Rortex/vortex identification method

Abstract: In this paper, a modified normalized Rortex/vortex identification method named ΩR is presented to improve the original ΩR method and resolve the bulging phenomenon on the isosurfaces, which is caused by the original ΩR method. Mathematical explanations and the relationship between the Q criterion and ΩR are described in detail. In addition, the new developed formula does not require two original coordinate rotations, and the calculation of ΩR is greatly simplified. The numerical results demonstrate the effectiveness of the new modified normalized Rortex/vortex identification method.In this paper, a modified normalized Rortex/vortex identification method named ΩR is presented to improve the original ΩR method and resolve the bulging phenomenon on the isosurfaces, which is caused by the original ΩR method. Mathematical explanations and the relationship between the Q criterion and ΩR are described in detail. In addition, the new developed formula does not require two original coordinate rotations, and the calculation of ΩR is greatly simplified. The numerical results demonstrate the effectiveness of the new modified normalized Rortex/vortex identification method.

Meltblown technology for production of polymeric microfibers/nanofibers: A review

Abstract: This work summarizes the current state of knowledge in the area of meltblown technology for production of polymeric nonwovens with specific attention to utilized polymers, die design, production of nanofibers, the effect of process variables (such as the throughput rate, melt rheology, melt temperature, die temperature, air temperature/velocity/pressure, die-to-collector distance, and speed) with relation to nonwoven characteristics as well as to typical flow instabilities such as whipping, die drool, fiber breakup, melt spraying, flies, generation of small isolated spherical particles, shots, jam, and generation of nonuniform fiber diameters.

Laser induced cavitation: Plasma generation and breakdown shockwave

Abstract: Laser induced cavitation is one of the effective techniques to generate controlled cavitation bubbles, both for basic study and for applications in different fields of engineering and medicine. Unfortunately, control of bubble formation and symmetry is hardly achieved due to a series of concurrent causes. In particular, the need to focus the laser beam at the bubble formation spot leads, in general, to a conical region proximal to the light source where conditions are met for plasma breakdown. A finite sized region then exists where the electric field may fluctuate depending on several disturbing agents, leading to possible plasma fragmentation and plasma intensity variation. Such irregularities may induce asymmetry in the successive bubble dynamics, a mostly undesired effect if reproducible conditions are sought for. In the present paper, the structure of the breakdown plasma and the ensuing bubble dynamics are analyzed by means of high speed imaging and intensity measurements of the shockwave system launched at breakdown. It is found that the parameters of the system can be tuned to optimize repeatability and sphericity. In particular, symmetric rebound dynamics is achieved almost deterministically when a pointlike plasma is generated at the breakdown threshold energy. Spherical symmetry is also favored by a large focusing angle combined with a relatively large pulse energy, a process which, however, retains a significant level of stochasticity. Outside these special conditions, the elongated and often fragmented conical plasma shape is found to be correlated with anisotropic and multiple breakdown shockwave emission.

Dynamic behavior of water drops impacting on cylindrical superhydrophobic surfaces

Abstract: This paper investigates the impact behavior between water drops with different velocities and cylindrical superhydrophobic surfaces with various diameters and presents two possible outcomes of drop impact, which are asymmetric rebound and stretched breakup. Due to the special cylindrical topology of the surface, drops undergo an asymmetric spreading and retracting process in the azimuthal and the axial direction, which results in three types of asymmetric rebound, including jug-like rebound, wing-like rebound, and rebound breakup. The stretched breakup is observed in the collision of drops with higher impact velocities and smaller cylinder diameters. The diameter ratio D* and Weber number We are found to be the determinants of the bouncing patterns. With the decrease in the diameter ratio D* or the increase in the Weber number We, the bouncing patterns transformed from jug-like rebound through wing-like rebound and finally to stretched breakup. We put forward a modification form of the Weber number (α = We/D*) affected by the diameter ratio D*, indicating the ratio between the inertia force and the surface tension, as the criterion to distinguish the upward rebound from the downward stretch, which helps obtain the linear relation of critical Wecr and D*cr. Furthermore, asymmetric rebound and stretched breakup could effectively shorten the contact time between drops and substrates. The contact time is found to be mainly determined by the dimensionless parameter α. The correlation between the dimensionless contact time and the dimensionless parameter α is demonstrated to be τc ∝ αn.

Dispersion due to combined pressure-driven and electro-osmotic flows in a channel surrounded by a permeable porous medium

Abstract: The combined electro-osmotic and pressure-driven flows (PDFs) have pronounced impacts on the solute transport in permeable porous media, particularly mixing and separation processes. However, the relationship between the physical properties of the permeable porous media and the combined electro-osmotic and PDFs still needs further investigation. This study focuses on the transport of a neutral nonreacting solute in a channel with permeable porous walls under the combined effects of electro-osmotic and PDFs. With the aid of perturbation theory and asymptotic analysis, the equivalent one-dimensional equations governing the solute concentrations in the channel and permeable porous medium under the combined velocity are derived. Based on this, an exact analytical expression relating the dispersion coefficient with the physical properties of the permeable porous medium and the combined flow is obtained. The model parameters exerting the most influence on the results are identified through sensitivity analysis. The proposed model is compared and validated with several previously developed models in the literature. The findings of this study can pave the way for the quantitatively design of solute transport through microporous coatings and porous microfluidic membranes.

Modeling realistic multiphase flows using a non-orthogonal multiple-relaxation-time lattice Boltzmann method

Abstract: In this paper, we develop a three-dimensional multiple-relaxation-time lattice Boltzmann method (MRT-LBM) based on a set of non-orthogonal basis vectors. Compared with the classical MRT-LBM based on a set of orthogonal basis vectors, the present non-orthogonal MRT-LBM simplifies the transformation between the discrete velocity space and the moment space and exhibits better portability across different lattices. The proposed method is then extended to multiphase flows at large density ratio with tunable surface tension, and its numerical stability and accuracy are well demonstrated by some benchmark cases. Using the proposed method, a practical case of a fuel droplet impacting on a dry surface at high Reynolds and Weber numbers is simulated and the evolution of the spreading film diameter agrees well with the experimental data. Furthermore, another realistic case of a droplet impacting on a super-hydrophobic wall with a cylindrical obstacle is reproduced, which confirms the experimental finding of Liu et al. [“Symmetry breaking in drop bouncing on curved surfaces,” Nat. Commun. 6, 10034 (2015)] that the contact time is minimized when the cylinder radius is comparable with the droplet radius.

Direct molecular simulation of internal energy relaxation and dissociation in oxygen

Abstract: A variant of the direct simulation Monte Carlo (DSMC) method, referred to as direct molecular simulation (DMS), is used to study oxygen dissociation from first principles. The sole model input to the DMS calculations consists of 12 potential energy surfaces that govern O2 + O2 and O + O2 collisions, including all spin-spatial degenerate configurations, in the ground electronic state. DMS calculations are representative of the gas evolution behind a strong shock wave, where molecular oxygen excites rotationally and vibrationally before ultimately dissociating and reaching a quasi-steady-state (QSS). Vibrational relaxation time constants are presented for both O2 + O2 and O + O2 collisions and are found to agree closely with experimental data. Compared to O2 + O2 collisions, vibrational relaxation due to O + O2 collisions is found to be ten times faster and to have a weak dependence on temperature. Dissociation rate constants in the QSS dissociation phase are presented for both O2 + O2 and O + O2 collisions and agree (within experimental uncertainty) with rates inferred from shock-tube experiments. Both experiments and simulations indicate that the QSS dissociation rate coefficients for O + O2 interactions are about two times greater than the ones for O2 + O2. DMS calculations predict this to be a result of nonequilibrium (non-Boltzmann) internal energy distributions. Specifically, the increased dissociation rate is caused by faster vibrational relaxation, due to O + O2 collisions, which alters the vibrational energy distribution function in the QSS by populating higher energy states that readily dissociate. Although existing experimental data appear to support this prediction, experiments with lower uncertainty are needed for quantitative validation. The DMS data presented for rovibrational relaxation and dissociation in oxygen could be used to formulate models for DSMC and computational fluid dynamics methods.A variant of the direct simulation Monte Carlo (DSMC) method, referred to as direct molecular simulation (DMS), is used to study oxygen dissociation from first principles. The sole model input to the DMS calculations consists of 12 potential energy surfaces that govern O2 + O2 and O + O2 collisions, including all spin-spatial degenerate configurations, in the ground electronic state. DMS calculations are representative of the gas evolution behind a strong shock wave, where molecular oxygen excites rotationally and vibrationally before ultimately dissociating and reaching a quasi-steady-state (QSS). Vibrational relaxation time constants are presented for both O2 + O2 and O + O2 collisions and are found to agree closely with experimental data. Compared to O2 + O2 collisions, vibrational relaxation due to O + O2 collisions is found to be ten times faster and to have a weak dependence on temperature. Dissociation rate constants in the QSS dissociation phase are presented for both O2 + O2 and O + O2 collisions...

MHD mixed convection and entropy generation of nanofluid in a lid-driven U-shaped cavity with internal heat and partial slip

Abstract: This contribution simulates the impact of partial slip on entropy generation due to magnetohydrodynamic, mixed convection of nanofluids in a lid-driven U-shaped cavity with discrete heating. The influence of the partial slip effect is proposed along the lid-driven vertical walls. A uniform heat flux source on the bottom wall is proposed; meanwhile, the two portions of the outer-upper horizontal walls are cooled isothermally. The remainder cavity walls are taken adiabatic. The governing equations are solved using the finite volume approach, and the outcomes are successfully validated against previous studies. Simulation results are presented and discussed for several cases with the impacts of the governing parameters on the heat transfer rate. Inspection of the results in mixed convective and entropy generation environments demonstrate that the average Nusselt number increases with the increase in the volume fraction of nanoparticles at AR = 0.1. For all values of D (heat source location), the Nusselt number increases by crossing the heat source and reaches its maximum value at the end of the source. Also, for all values of the aspect ratio, addition of nanoparticles into the base fluid leads to a loss in the thermal performance. Moreover, for all states of movement, addition of nanoparticles into the base fluid leads to an increase in the entropy.

The suction effect during freak wave slamming on a fixed platform deck: Smoothed particle hydrodynamics simulation and experimental study

Abstract: During the process of wave slamming on a structure with sharp corners, the wave receding after wave impingement can induce strong negative pressure (relative to the atmospheric pressure) at the bottom of the structure, which is called the suction effect. From the practical point of view, the suction force induced by the negative pressure, coinciding with the gravity force, pulls the structure down and hence increases the risk of structural damage. In this work, the smoothed particle hydrodynamics (SPH) method, more specifically the δ+SPH model, is adopted to simulate the freak wave slamming on a fixed platform with the consideration of the suction effect, i.e., negative pressure, which is a challenging issue because it can cause the so-called tensile instability in SPH simulations. The key to overcome the numerical issue is to use a numerical technique named tensile instability control (TIC). Comparative studies using SPH models with and without TIC will show the importance of this technique in capturing the negative pressure. It is also found that using a two-phase simulation that takes the air phase into account is essential for an SPH model to accurately predict the impact pressure during the initial slamming stage. The freak wave impacts with different water depths are studied. All the multiphase SPH results are validated by our experimental data. The wave kinematics/dynamics and wave impact features in the wave-structure interacting process are discussed, and the mechanism of the suction effect characterized by the negative pressure is carefully analyzed.

Turbulent drag reduction by polymer additives: Fundamentals and recent advances

Abstract: (NSERC) Natural Sciences and Engineering Research Council of Canada (No. RGPIN-2014-04903); (NSF) National Science Foundation (No. NSF PHY11-25915)

Active control of vortex-induced vibration of a circular cylinder using machine learning

Abstract: We demonstrate the use of high-fidelity computational fluid dynamics simulations in machine-learning based active flow control. More specifically, for the first time, we adopt the genetic programming (GP) to select explicit control laws, in a data-driven and unsupervised manner, for the suppression of vortex-induced vibration (VIV) of a circular cylinder in a low-Reynolds-number flow (Re = 100), using blowing/suction at fixed locations. A cost function that balances both VIV suppression and energy consumption for the control is carefully chosen according to the knowledge obtained from pure blowing/suction open-loop controls. By implementing reasonable constraints to VIV amplitude and actuation strength during the GP evolution, the GP-selected best ten control laws all point to suction-type actuation. The best control law suggests that the suction strength should be nonzero when the cylinder is at its equilibrium position and should increase nonlinearly with the cylinder’s transverse displacement. Applying this control law suppresses 94.2% of the VIV amplitude and achieves 21.4% better overall performance than the best open-loop controls. Furthermore, it is found that the GP-selected control law is robust, being effective in flows ranging from Re = 100 to 400. On the contrary, although the P-control can achieve similar performance as the GP-selected control at Re = 100, it deteriorates in higher Reynolds number flows. Although for demonstration purpose the chosen control problem is relatively simple, the training experience and insights obtained from this study can shed some light on future GP-based control of more complicated problems.

Self-driven droplet transport: Effect of wettability gradient and confinement

Abstract: Surface tension driven droplet transport in an open surface is increasingly becoming popular for various microfluidic applications. In this work, efficient transport of a glycerin droplet on an open wettability gradient surface with controlled wettability and confinement is numerically investigated. Nondimensional track width w* (ratio of the width of the wettability gradient track w and the initial droplet diameter d0) of a wettability gradient track laid on a superhydrophobic background represents wettability confinement. A lower value of w* represents higher wettability confinement. Droplet behavior changes for different wettability confinements and gradients of the track. It is found that droplet velocity is a function of the wettability confinement and the gradient; droplet transport velocity is maximum for w* = 0.8. Higher confinement (w* 0.6 irrespective of the wettability gradient of the track. These findings provide valuable insight into efficient droplet manipulation in microfluidic devices.

Modeling the interplay between the shear layer and leading edge suction during dynamic stall

Abstract: The dynamic stall development on a pitching airfoil at Re = 106 was investigated by time-resolved surface pressure and velocity field measurements. Two stages were identified in the dynamic stall development based on the shear layer evolution. In the first stage, the flow detaches from the trailing edge and the separation point moves gradually upstream. The second stage is characterized by the roll up of the shear layer into a large scale dynamic stall vortex. The two-stage dynamic stall development was independently confirmed by global velocity field and local surface pressure measurements around the leading edge. The leading edge pressure signals were combined into a single leading edge suction parameter. We developed an improved model of the leading edge suction parameter based on thin airfoil theory that links the evolution of the leading edge suction and the shear layer growth during stall development. The shear layer development leads to a change in the effective camber and the effective angle of attack. By taking into account this twofold influence, the model accurately predicts the value and timing of the maximum leading edge suction on a pitching airfoil. The evolution of the experimentally obtained leading edge suction was further analyzed for various sinusoidal motions revealing an increase in the critical value of the leading edge suction parameter with increasing pitch unsteadiness. The characteristic dynamic stall delay decreases with increasing unsteadiness, and the dynamic stall onset is best assessed by critical values of the circulation and the shear layer height which are motion independent.