SAM
https://sam.ensam.eu:443
The DSpace digital repository system captures, stores, indexes, preserves, and distributes digital research material.Tue, 21 May 2019 07:44:13 GMT2019-05-21T07:44:13ZNumerical investigation of the real and ideal gap profiles in the calculation of the pressure distortion coefficient and piston fall rate of an LNE 200 MPa pressure balance
http://hdl.handle.net/10985/8914
Numerical investigation of the real and ideal gap profiles in the calculation of the pressure distortion coefficient and piston fall rate of an LNE 200 MPa pressure balance
WONGTHEP, Padipat; RABAULT, Thierry; NOGUERA, Ricardo; SARRAF, Christophe
This paper aims to investigate, by means of numerical simulation, the effect of gap profiles on the calculation of the pressure distortion coefficient (λ) and the piston fall rate (vf) of two piston-cylinder units used in a Laboratoire National de Métrologie et d'Essais (LNE) 200 MPa pressure balance. The ideal mean gap width between the piston and the cylinder was obtained after measuring the piston fall rate at a low pressure, while the piston radius was obtained from the cross-float experiments at a low pressure. The real gap width was obtained from dimensional measurements by measuring the diameter and straightness of the piston and the cylinder. The piston and cylinder radial distortions were calculated using the finite element method. The pressure distribution in the gap was calculated on the basis of the Navier-Stokes equation for Newtonian viscous flow. The results such as pressure distributions, radial distortions, the pressure distortion coefficient and the piston fall rate were presented for the free-deformation operating mode of the assemblies. The calculation resulted in ideal and real gap profiles indicating that the average pressure distortion coefficient was in good agreement within 0.017 × 10-6 MPa-1 and the calculations of piston fall rate depended on the gap profile especially at the inlet and outlet zones of the engagement length.
Tue, 01 Jan 2013 00:00:00 GMThttp://hdl.handle.net/10985/89142013-01-01T00:00:00ZWONGTHEP, PadipatRABAULT, ThierryNOGUERA, RicardoSARRAF, ChristopheThis paper aims to investigate, by means of numerical simulation, the effect of gap profiles on the calculation of the pressure distortion coefficient (λ) and the piston fall rate (vf) of two piston-cylinder units used in a Laboratoire National de Métrologie et d'Essais (LNE) 200 MPa pressure balance. The ideal mean gap width between the piston and the cylinder was obtained after measuring the piston fall rate at a low pressure, while the piston radius was obtained from the cross-float experiments at a low pressure. The real gap width was obtained from dimensional measurements by measuring the diameter and straightness of the piston and the cylinder. The piston and cylinder radial distortions were calculated using the finite element method. The pressure distribution in the gap was calculated on the basis of the Navier-Stokes equation for Newtonian viscous flow. The results such as pressure distributions, radial distortions, the pressure distortion coefficient and the piston fall rate were presented for the free-deformation operating mode of the assemblies. The calculation resulted in ideal and real gap profiles indicating that the average pressure distortion coefficient was in good agreement within 0.017 × 10-6 MPa-1 and the calculations of piston fall rate depended on the gap profile especially at the inlet and outlet zones of the engagement length.Experimental study of blade rigidity effects on the global and the local performances of a thick blades axial-flow fan
http://hdl.handle.net/10985/8584
Experimental study of blade rigidity effects on the global and the local performances of a thick blades axial-flow fan
NOURI, Hussain; RAVELET, Florent; SARRAF, Christophe; BAKIR, Farid
An experimental investigation on the aerodynamic performances of thick blades axial-flow fans was carried out in this study. Two fans are considered, the first one is rotomoulded (in plastic) and the second one is milled (in aluminium). Both have exactly the same shape, excepting that the rotomoulded fan has hollow blades. They were designed from an existing fan (manufactured by plastic injection process) used in the cooling system of an automotive vehicle power unit. As far as shape is concerned, the only difference between the two first fans and the traditional injected fan is the blade thickness, whereas as far as rigidity is concerned, the only difference between the rotomoulded and the milled fans is the ability of the rotomoulded fan to be deformed easier than the milled fan. The aim of this study is to determine on the one hand the influence of the blade thickness and on the other hand the way the deformation of the hollow blades may affect the global and the local performances. The global performances of the fans were measured in a test bench designed according to the ISO 5801 standards. The curve of the aerodynamics characteristics (pressure head versus flow rate) and of the global efficiency are slightly lower for the rotomoulded fan. The wall pressure fluctuations were also investigated for three flow rates: one corresponding to the maximum efficiencies of both fans and two others corresponding to an under-flow and an over-flow rate. The power spectral density (PSD) levels, estimated by the Welch method, are between six and nine times higher for the rotomoulded fan at nominal flow rate. At partial flow rate, however, the PSD levels are close for both fans.
Fri, 01 Jan 2010 00:00:00 GMThttp://hdl.handle.net/10985/85842010-01-01T00:00:00ZNOURI, HussainRAVELET, FlorentSARRAF, ChristopheBAKIR, FaridAn experimental investigation on the aerodynamic performances of thick blades axial-flow fans was carried out in this study. Two fans are considered, the first one is rotomoulded (in plastic) and the second one is milled (in aluminium). Both have exactly the same shape, excepting that the rotomoulded fan has hollow blades. They were designed from an existing fan (manufactured by plastic injection process) used in the cooling system of an automotive vehicle power unit. As far as shape is concerned, the only difference between the two first fans and the traditional injected fan is the blade thickness, whereas as far as rigidity is concerned, the only difference between the rotomoulded and the milled fans is the ability of the rotomoulded fan to be deformed easier than the milled fan. The aim of this study is to determine on the one hand the influence of the blade thickness and on the other hand the way the deformation of the hollow blades may affect the global and the local performances. The global performances of the fans were measured in a test bench designed according to the ISO 5801 standards. The curve of the aerodynamics characteristics (pressure head versus flow rate) and of the global efficiency are slightly lower for the rotomoulded fan. The wall pressure fluctuations were also investigated for three flow rates: one corresponding to the maximum efficiencies of both fans and two others corresponding to an under-flow and an over-flow rate. The power spectral density (PSD) levels, estimated by the Welch method, are between six and nine times higher for the rotomoulded fan at nominal flow rate. At partial flow rate, however, the PSD levels are close for both fans.Experimental study of blade thickness effects on the overall and local performances of a Controlled Vortex Designed axial-ﬂow fan
http://hdl.handle.net/10985/11220
Experimental study of blade thickness effects on the overall and local performances of a Controlled Vortex Designed axial-ﬂow fan
SARRAF, Christophe; NOURI, Hussain; RAVELET, Florent; BAKIR, Farid
The purpose of this work is to study the effects of blade thickness on the performances of an axial-ﬂow fan. Two fans that differ only in the thickness of their blades were studied. The ﬁrst fan was designed to be part of the cooling system of an automotive vehicle power unit and has very thin blades. The second fan has much thicker blades compatible with the rotomoulding conception process. The overall perfor-mances of the fans were measured in a test bench designed according to the ISO-5801 standard. The curve of aerodynamics characteristics (pressure head versus ﬂow-rate) is slightly steeper for the fan with thick blades, and the nominal point is shifted towards lower ﬂow-rates. The efﬁciency of the thick blades fan is lower than the efﬁciency of the fan with thin blades but remains high on a wider ﬂow-rate range. The mean velocity ﬁelds downstream of the rotors are very similar at nominal points with less centrifu-gation for the thick blades fan. Moreover, the thick blades fan maintains an axial exit-ﬂow on a wider range of ﬂow-rates. The main differences concern local properties of the ﬂow: phase-averaged velocities and wall pressure ﬂuctuations strongly differ at the nominal ﬂow-rates. The total level of ﬂuctuations is lower for the thick blades fan that for the thin blades fan and the spectral decomposition of the wall ﬂuc-tuations and velocity signals reveal more harmonics for the thick blades fan, with less correlation between the different signals. For this kind of turbomachinery, the use of thick blades could lead to a good compromise between aerodynamic and acoustic performances, on a wider operating range.
Sat, 01 Jan 2011 00:00:00 GMThttp://hdl.handle.net/10985/112202011-01-01T00:00:00ZSARRAF, ChristopheNOURI, HussainRAVELET, FlorentBAKIR, FaridThe purpose of this work is to study the effects of blade thickness on the performances of an axial-ﬂow fan. Two fans that differ only in the thickness of their blades were studied. The ﬁrst fan was designed to be part of the cooling system of an automotive vehicle power unit and has very thin blades. The second fan has much thicker blades compatible with the rotomoulding conception process. The overall perfor-mances of the fans were measured in a test bench designed according to the ISO-5801 standard. The curve of aerodynamics characteristics (pressure head versus ﬂow-rate) is slightly steeper for the fan with thick blades, and the nominal point is shifted towards lower ﬂow-rates. The efﬁciency of the thick blades fan is lower than the efﬁciency of the fan with thin blades but remains high on a wider ﬂow-rate range. The mean velocity ﬁelds downstream of the rotors are very similar at nominal points with less centrifu-gation for the thick blades fan. Moreover, the thick blades fan maintains an axial exit-ﬂow on a wider range of ﬂow-rates. The main differences concern local properties of the ﬂow: phase-averaged velocities and wall pressure ﬂuctuations strongly differ at the nominal ﬂow-rates. The total level of ﬂuctuations is lower for the thick blades fan that for the thin blades fan and the spectral decomposition of the wall ﬂuc-tuations and velocity signals reveal more harmonics for the thick blades fan, with less correlation between the different signals. For this kind of turbomachinery, the use of thick blades could lead to a good compromise between aerodynamic and acoustic performances, on a wider operating range.A new model of fluid flow to determine pressure balance characteristics
http://hdl.handle.net/10985/8900
A new model of fluid flow to determine pressure balance characteristics
WONGTHEP, Padipat; RABAULT, Thierry; NOGUERA, Ricardo; SARRAF, Christophe
Some projects such as the EUROMET project 463 have underlined the lack of agreement between experimental measurements and calculations by the finite element method (FEM), used to determine the piston fall rate of a high-pressure balance used in primary standards. This is significant because the piston fall rate is an essential parameter to characterize experimentally the mean gap between the piston and the cylinder and to determine the effective area (A p) at each pressure (p) point. By improving the method used to estimate the piston fall rate it is possible to improve the determination of the gap, the effective area and consequently the pressure distortion coefficient. One possible cause of the lack of agreement between the calculated and measured piston fall rates could be inappropriate modelling of the fluid flow. In fact, the former quasi-1D Stokes model assimilates the gap between the piston and the cylinder as formed by two parallel walls, which is an approximation. In addition, the velocity of the piston wall was neglected. In order to evaluate the influence of this model, the equations of the fluid flow are modified and are presented in this paper. Equations that were defined in a parallel-plane model are defined in an annular gap model. In addition to this, corrections due to the velocity of the piston wall are inserted. This research work is applied on a Desgranges et Huot DH 7594 piston-cylinder unit of PTB with a pressure up to 1 GPa, in the continuity of the EUROMET project 463 in order to quantify the influence of each correction that has been inserted in the new equations. This is carried out using the FEM. This analysis will allow us to evaluate the improvement of our knowledge of the behaviour of piston gauges and consequently to better evaluate the uncertainties due to the models.
Tue, 01 Jan 2013 00:00:00 GMThttp://hdl.handle.net/10985/89002013-01-01T00:00:00ZWONGTHEP, PadipatRABAULT, ThierryNOGUERA, RicardoSARRAF, ChristopheSome projects such as the EUROMET project 463 have underlined the lack of agreement between experimental measurements and calculations by the finite element method (FEM), used to determine the piston fall rate of a high-pressure balance used in primary standards. This is significant because the piston fall rate is an essential parameter to characterize experimentally the mean gap between the piston and the cylinder and to determine the effective area (A p) at each pressure (p) point. By improving the method used to estimate the piston fall rate it is possible to improve the determination of the gap, the effective area and consequently the pressure distortion coefficient. One possible cause of the lack of agreement between the calculated and measured piston fall rates could be inappropriate modelling of the fluid flow. In fact, the former quasi-1D Stokes model assimilates the gap between the piston and the cylinder as formed by two parallel walls, which is an approximation. In addition, the velocity of the piston wall was neglected. In order to evaluate the influence of this model, the equations of the fluid flow are modified and are presented in this paper. Equations that were defined in a parallel-plane model are defined in an annular gap model. In addition to this, corrections due to the velocity of the piston wall are inserted. This research work is applied on a Desgranges et Huot DH 7594 piston-cylinder unit of PTB with a pressure up to 1 GPa, in the continuity of the EUROMET project 463 in order to quantify the influence of each correction that has been inserted in the new equations. This is carried out using the FEM. This analysis will allow us to evaluate the improvement of our knowledge of the behaviour of piston gauges and consequently to better evaluate the uncertainties due to the models.Experimental study of the instationary flow between two ducted Counter-rotating rotors
http://hdl.handle.net/10985/8536
Experimental study of the instationary flow between two ducted Counter-rotating rotors
NOURI, Hussain; DANLOS, Amélie; RAVELET, Florent; BAKIR, Farid; SARRAF, Christophe
The purpose of this work is to study experimentally the aerodynamic characteristics of a subsonic counter-rotating axial-flow fans system operating in a ducted configuration. The fans of diameter D = 375 mm were designed to match the specification point using an original iterative method: the front rotor blade cascade is designed with a conventional inverse method, setting the radial distribution of the Euler work. The through-flow is then computed using an axisymmetric and radial equilibrium asumption, with empirical models of losses. The rear rotor is not conventional but is designed to straighten the radial profile of the tangential velocity. The design of the front rotor is then modified until the stage meets the requirements. The experimental setup is arranged such that the rotation rate of each fan is independently controlled and that the axial distance between the rotors can be varied from 17% to 310% of the mid-span chord length. Systematic measurements of the global performances and local measurements of the velocity field and of the wall pressure fluctuations are performed, in order to first validate the design method, and to explore the effects of the two specific free parameters of the system: the axial spacing and the ratio of rotation rates. The results show that the efficiency is strongly increased compared to a conventional rotor or to a rotor-stator stage. The developed design method slightly over-predicts the pressure rise and slightly under-predicts the best ratio of rotation rates. Flow angle measurements downstream of the stage show that the outflow is not completely straightened at the design point. Finally, the system is highly efficient on a wide range of flow-rates and pressure rises: this system has thus a very flexible use, with a large patch of high efficient operating points in the parameter space.
Tue, 01 Jan 2013 00:00:00 GMThttp://hdl.handle.net/10985/85362013-01-01T00:00:00ZNOURI, HussainDANLOS, AmélieRAVELET, FlorentBAKIR, FaridSARRAF, ChristopheThe purpose of this work is to study experimentally the aerodynamic characteristics of a subsonic counter-rotating axial-flow fans system operating in a ducted configuration. The fans of diameter D = 375 mm were designed to match the specification point using an original iterative method: the front rotor blade cascade is designed with a conventional inverse method, setting the radial distribution of the Euler work. The through-flow is then computed using an axisymmetric and radial equilibrium asumption, with empirical models of losses. The rear rotor is not conventional but is designed to straighten the radial profile of the tangential velocity. The design of the front rotor is then modified until the stage meets the requirements. The experimental setup is arranged such that the rotation rate of each fan is independently controlled and that the axial distance between the rotors can be varied from 17% to 310% of the mid-span chord length. Systematic measurements of the global performances and local measurements of the velocity field and of the wall pressure fluctuations are performed, in order to first validate the design method, and to explore the effects of the two specific free parameters of the system: the axial spacing and the ratio of rotation rates. The results show that the efficiency is strongly increased compared to a conventional rotor or to a rotor-stator stage. The developed design method slightly over-predicts the pressure rise and slightly under-predicts the best ratio of rotation rates. Flow angle measurements downstream of the stage show that the outflow is not completely straightened at the design point. Finally, the system is highly efficient on a wide range of flow-rates and pressure rises: this system has thus a very flexible use, with a large patch of high efficient operating points in the parameter space.Critical evaluation of three hemodynamic models for the numerical simulation of intra-stent flows
http://hdl.handle.net/10985/9961
Critical evaluation of three hemodynamic models for the numerical simulation of intra-stent flows
CHABI, Fatiha; CHAMPMARTIN, Stéphane; SARRAF, Christophe; NOGUERA, Ricardo
We evaluate here three hemodynamic models used for the numerical simulation of bare and stented artery flows. We focus on two flow features responsible for intra-stent restenosis: the wall shear stress and the re-circulation lengths around a stent. The studied models are the Poiseuille profile, the simplified pulsatile profile and the complete pulsatile profile based on the analysis of Womersley. The flow rate of blood in a human left coronary artery is considered to compute the velocity profiles. “Ansys Fluent 14.5” is used to solve the Navier–Stokes and continuity equations. As expected our results show that the Poiseuille profile is questionable to simulate the complex flow dynamics involved in intra-stent restenosis. Both pulsatile models give similar results close to the strut but diverge far from it. However, the computational time for the complete pulsatile model is five times that of the simplified pulsatile model. Considering the additional “cost” for the complete model, we recommend using the simplified pulsatile model for future intra-stent flow simulations.
Thu, 01 Jan 2015 00:00:00 GMThttp://hdl.handle.net/10985/99612015-01-01T00:00:00ZCHABI, FatihaCHAMPMARTIN, StéphaneSARRAF, ChristopheNOGUERA, RicardoWe evaluate here three hemodynamic models used for the numerical simulation of bare and stented artery flows. We focus on two flow features responsible for intra-stent restenosis: the wall shear stress and the re-circulation lengths around a stent. The studied models are the Poiseuille profile, the simplified pulsatile profile and the complete pulsatile profile based on the analysis of Womersley. The flow rate of blood in a human left coronary artery is considered to compute the velocity profiles. “Ansys Fluent 14.5” is used to solve the Navier–Stokes and continuity equations. As expected our results show that the Poiseuille profile is questionable to simulate the complex flow dynamics involved in intra-stent restenosis. Both pulsatile models give similar results close to the strut but diverge far from it. However, the computational time for the complete pulsatile model is five times that of the simplified pulsatile model. Considering the additional “cost” for the complete model, we recommend using the simplified pulsatile model for future intra-stent flow simulations.Experimental study of aerated cavitation in a horizontal venturi nozzle
http://hdl.handle.net/10985/9882
Experimental study of aerated cavitation in a horizontal venturi nozzle
TOMOV, Petar; KHELLADI, Sofiane; RAVELET, Florent; SARRAF, Christophe; BAKIR, Farid; VERTENOEUIL, P.
The injection of bubbles into an already cavitating flow is a way of influencing the typical cavitating behavior. The present article deals with experiments on aerated and non-aerated cavitation in a transparent horizontal venturi nozzle. The observations are done by means of a high-speed camera. In such a way the extremely rapid cavitation and cavitation-aeration flows are captured and further analysed. The post-processing techniques is based on the detection of the grey level on the series of images. As a result, three different regimes are identified: sheet cavitation, cloud cavitation and “supercavitation”. Those regimes are further aerated by injecting air bubbles. Standard deviations, time- space diagrams and frequency spectrum based on the vertical distribution of the grey level along a monitored line are plotted for all of the observed regimes. In the pure cavitation cases we obtain statistically symmetrical structures with characteristic lengths and frequencies. On the other hand, with aeration present, the symmetry is broken and characteristic lengths and frequencies are deeply modified, until a complete disappearance when “supercavitation” is reached.
Thu, 01 Jan 2015 00:00:00 GMThttp://hdl.handle.net/10985/98822015-01-01T00:00:00ZTOMOV, PetarKHELLADI, SofianeRAVELET, FlorentSARRAF, ChristopheBAKIR, FaridVERTENOEUIL, P.The injection of bubbles into an already cavitating flow is a way of influencing the typical cavitating behavior. The present article deals with experiments on aerated and non-aerated cavitation in a transparent horizontal venturi nozzle. The observations are done by means of a high-speed camera. In such a way the extremely rapid cavitation and cavitation-aeration flows are captured and further analysed. The post-processing techniques is based on the detection of the grey level on the series of images. As a result, three different regimes are identified: sheet cavitation, cloud cavitation and “supercavitation”. Those regimes are further aerated by injecting air bubbles. Standard deviations, time- space diagrams and frequency spectrum based on the vertical distribution of the grey level along a monitored line are plotted for all of the observed regimes. In the pure cavitation cases we obtain statistically symmetrical structures with characteristic lengths and frequencies. On the other hand, with aeration present, the symmetry is broken and characteristic lengths and frequencies are deeply modified, until a complete disappearance when “supercavitation” is reached.POD study of aerated cavitation in a venturi nozzle
http://hdl.handle.net/10985/10296
POD study of aerated cavitation in a venturi nozzle
TOMOV, Petar; DANLOS, Amélie; KHELLADI, Sofiane; RAVELET, Florent; SARRAF, Christophe; BAKIR, Farid
The fact of injecting bubbles into a cavitating flow influences typical cavitating behavior. Cavitation and aerated cavitation experiments has been carried out on a symmetrical venturi nozzle with convergent/divergent angles of 18 ◦ and 8 ◦ , respectively. A snapshot Proper Orthogonal Decomposition (POD) technique is used to identify different modes in terms of discharge flow velocity, pressure and injected quantity of air. The energy spectrum per given mode is also presented. The first four modes are outlined in the present paper for an aerated and non-aerated cavitating flows.
Thu, 01 Jan 2015 00:00:00 GMThttp://hdl.handle.net/10985/102962015-01-01T00:00:00ZTOMOV, PetarDANLOS, AmélieKHELLADI, SofianeRAVELET, FlorentSARRAF, ChristopheBAKIR, FaridThe fact of injecting bubbles into a cavitating flow influences typical cavitating behavior. Cavitation and aerated cavitation experiments has been carried out on a symmetrical venturi nozzle with convergent/divergent angles of 18 ◦ and 8 ◦ , respectively. A snapshot Proper Orthogonal Decomposition (POD) technique is used to identify different modes in terms of discharge flow velocity, pressure and injected quantity of air. The energy spectrum per given mode is also presented. The first four modes are outlined in the present paper for an aerated and non-aerated cavitating flows.Study of passive control study of the cavitation instability on a venturi profile
http://hdl.handle.net/10985/11254
Study of passive control study of the cavitation instability on a venturi profile
DANLOS, Amélie; MEHAL, Jean-Elie; SARRAF, Christophe; RAVELET, Florent
This paper presents experimental and numerical results concerning the operation of a configuration that includes an axial pump and a bundle of tubes that mimics the cool source of a heat exchanger. The pump used in the tests has a low solidity and two blades designed in forced vortex, the tip clearance is approximately 3.87% of tip radius. The experimental measures of the characteristic curves and of the wall static pressure fluctuations signals upstream and downstream the pump and the exchanger were compared to the numerical results. The numerical simulations were carried out by using a Fluent code, with URANS (Unsteady Reynolds Averaged Navier-Stokes) approach and k-w SST turbulence model.
Fri, 01 Jan 2010 00:00:00 GMThttp://hdl.handle.net/10985/112542010-01-01T00:00:00ZDANLOS, AmélieMEHAL, Jean-ElieSARRAF, ChristopheRAVELET, FlorentThis paper presents experimental and numerical results concerning the operation of a configuration that includes an axial pump and a bundle of tubes that mimics the cool source of a heat exchanger. The pump used in the tests has a low solidity and two blades designed in forced vortex, the tip clearance is approximately 3.87% of tip radius. The experimental measures of the characteristic curves and of the wall static pressure fluctuations signals upstream and downstream the pump and the exchanger were compared to the numerical results. The numerical simulations were carried out by using a Fluent code, with URANS (Unsteady Reynolds Averaged Navier-Stokes) approach and k-w SST turbulence model.Thickness effect of NACA foils on hydrodynamic global parameters, boundary layer states and stall establishment
http://hdl.handle.net/10985/10295
Thickness effect of NACA foils on hydrodynamic global parameters, boundary layer states and stall establishment
SARRAF, Christophe; DJERIDI, Henda; PROTHIN, Sébastien; BILLARD, Jean-Yves
The present study investigates experimentally the hydrodynamic behavior of 2D NACA (15, 25, 35%) symmetric hydrofoils at Reynolds number 0.5 106 . A particular attention is paid on the hysteretic behavior at static stall angle and a detailed cartography of boundary layer structures (integral quantities and velocity profiles) is given in order to put in evidence the mechanism of the detachment and the onset of von Karman instability for thick profile
Fri, 01 Jan 2010 00:00:00 GMThttp://hdl.handle.net/10985/102952010-01-01T00:00:00ZSARRAF, ChristopheDJERIDI, HendaPROTHIN, SébastienBILLARD, Jean-YvesThe present study investigates experimentally the hydrodynamic behavior of 2D NACA (15, 25, 35%) symmetric hydrofoils at Reynolds number 0.5 106 . A particular attention is paid on the hysteretic behavior at static stall angle and a detailed cartography of boundary layer structures (integral quantities and velocity profiles) is given in order to put in evidence the mechanism of the detachment and the onset of von Karman instability for thick profile