A Research Methodology to Study Jet Wiping Processes

See discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/282851961 A Research Methodology to Study Jet Wiping Processes CONFERENCE PAPER · SEPTEMBER 2015 DOI: 10.13140/RG.2.1.1832.2641 READS 12 4 AUTHORS, INCLUDING: Miguel Alfonso Mendez Anne Gosset 3 PUBLICATIONS 1 CITATION 15 PUBLICATIONS 74 CITATIONS von Karman Institute for Fluid Dynamics SEE PROFILE University of A Coruña SEE PROFILE Available from: Miguel Alfonso Mendez Retrieved on: 24 October 2015 A Research Methodology to Study Jet Wiping Processes M.A. Mendez1 , K. Myrillas1 , A. Gosset2 and J.-M. Buchlin1 1 2 EA Dep., von Karman Institute for Fluid Dynamics, Sint-Genesius-Rode, Belgium Naval and Oceanic engineering Department, University of A Coruña, Ferrol, Spain Corresponding author: mendez@vki.ac.be Keywords: Jet Wiping, Falling Liquid Films, Impinging Gas Jets, Hydrodynamic Feedback In the jet wiping process a strip is withdrawn from a bath of coating liquid and impinged by a gas jet to control the final coating thickness (Fig. 1). The process is continuous, contact-less and simple. On the other hand, the achievable coating uniformity and the production rates are limited by the occurrence of undulation patterns on the final coating, which degrades the optical properties of the products. Despite being a major industrial stake, the causes of undulation remains poorly understood. The pioneer work of Tuck ([1]), recently extended by Hocking et al [2], has shown that the final coat is marginally stable and incapable of developing the undulation alone: an external perturbation is required. Several authors identify a buckling instability of the impinging jet (buckling hypothesis) or vibration of the steel strip (vibration hypothesis) as possible external sources of the undulation ([3-5]). Nevertheless, the underlying physics remains poorly understood and requires a systematic investigation. The methodology developed at the von Karman Institute (VKI) aims at identifying the physical mechanism responsible for coating undulation by studying the jet wiping process on a set of simplified scenarios. The undulation problem was first reproduced on a laboratory scale (Sec.1), then the gas-liquid interaction was studied in detail via flow visualization and numerical analysis (Sec.2). The identified destabilizing mechanism was then reproduced and tested on a simplified test case (Sec.3). Fig. 1: Sketch of a jet wiping process and example of typical undulation pattern. 1. Analysis of Undulation Patterns This step aimed at characterizing the film flow patterns as a function of the process parameters, testing several hypotheses on the undulation origin, and studying the frequency content of the gas jet and the undulation pattern on the liquid (Fig. 2). The facility consists of a large rotating cylinder which entrains a film of Dipropylene Glycol. The light absorption technique is used for film thickness characterization and hot wire anemometry is used for gas jet frequency analysis. Fig. 2: a) Experimental Set Up and b) Undulation pattern in terms of peak to peak amplitudes Ap and mean film thickness hf , as a function of stand-off distance Z and nozzle pressure Pn . The wavelength of the undulation ranged between λ=20-40mm, with amplitudes between 10-40% of the mean thickness hf . For a withdrawal speed of U = 0.34m/s, a typical frequency of the undulation ranged f =840Hz. As the jet buckling instability of the jet is supposedly occurring at much higher frequencies (200500kHz, [3]) , the buckling hypothesis was rejected. In order to test the vibration hypothesis, we observed that the wiping actuators (maximum pressure gradient ∇Pmax and maximum shear stress τmax ) are linearly dependent on the stand-off distance Z. Therefore, the effects of an oscillation of the strip (i.e. a standoff distance Z ± δZ oscillation) can be reproduced by nozzle pressure pulsation Pn ± δP . These pulsations were introduced by means of a rotating disk, which periodically opens and closes the supply line of the nozzle. Fig. 3 compares the spectra of the liquid film thicknesses impinged by a continuous jet and a pulsing jet. As no appreciable difference was observed for pressure oscillations corresponding to 0.15Z, the vibration hypothesis was rejected. Finally, a simultaneous characterization of air jet and liquid film flow revealed a strong matching of the spectra of the two flows, suggesting a jet flapping behavior occurring at the undulation frequency. To understand the physics of this mechanism, a more detailed analysis of the flow interaction was needed. Fig. 3: Spectrum of the liquid film impinged by a continuous jet (dashed black line) and by a pulsed jet (red line) at 5, 15, 30 Hz. The tests were performed with a nozzle with an opening d = 2mm, Pn = 165P a (Re ≈ 2300), Z/d = 8, U = 0.29m/s. The amplitude of the pressure pulsation is δP = 0.16Pn . 2. Analysis of Flows Interaction The second step aimed at visualizing the interaction between the two flows. First, the previous facility was equipped with a flow visualization set up: a high speed camera visualizing a cross section of the gas flow, which is seeded by oil particles and illuminated by a laser sheet (Fig. 4a). Second, the same wiping configuration was simulated numerically combining VOF for the gas-liquid interface and LES for the jet flow turbulence (Fig. 4b) [6]. Both approaches confirmed that the jet oscillates and prints its oscillation on the final coat. Moreover, the runback flow is pulsating, unsteadily confining the jet flow and producing flow recirculation below it. To understand whether one of the two flows is the triggering source of this coupling instability, a simplified configuration was needed. Fig. 4: Flow visualization and 3D VOF-LES simulation of a jet wiping process. 3. Simplified Test Case The third step aimed at de-coupling the instabilities of the two flows by artificially imposing one of them. To this end, a second laboratory model was designed (Fig 5): it consists of a slot air jet impinging on a liquid film falling along a fixed wall with a flow rate pulsing at controlled frequencies and amplitudes. The objective was to understand if a perturbation in the impinged liquid film propagates to the impinging jet, which possibly responds oscillating. The experimental facility for the liquid film perturbation is sketched in Fig. 6. The liquid flow in the test section is sustained by a hydrostatic and a sinusoidal gauge pressure, that is produced by two rotating butterfly valves which open and close the pneumatic lines pressurizing the circuit. The facility operates with Dipropylene Glycole, at mean Reynolds number ReM =q/ν =1-10, with q the flow rate per unit width and ν the liquid’s kinematic viscosity. A preliminary step consisted in characterizing the full set of perturbation reproducible, to ensure a full control of the gravity waves directed towards the jet impingement. Fig. 5: Sketch of the simplified test. Fig. 6: a) VKI Liquid film Facility, b) light absorption reconstruction of pulsing liquid film, c) liquid film wave peaks ĥM and toughs ĥm as a function of Re and fp . This facility was equipped with a light absorption test section. Fig. 6b shows an example of 3D thickness mapping, used to prove the bi-dimensionality of the perturbation. Fig. 6b shows the liquid film wave peaks ĥM and troughs ĥm , normalized with the mean thickness h, as a function of ReM , for different pulsation frequencies fp . The jet issues from a nozzle at Pn = 200P a, with a cross section 1.5 × 100mm2 , impinging 12cm below the liquid test section inlet. This nozzle is equipped with a laser sheet and two seeding injections for flow visualization: one traces the jet flow, the other traces the entrainment flow. A typical snapshot is in Fig. 7: the impinging air jet (1) deforms the liquid film asymmetrically into a 2D pulsing rib (2). This rib deflects a portion of the air flow, which forms, together with the entrainment flow, recirculation structures (3) above the jet. In the characterization of the jet flow (1), only the primary seeding injection is active and the video sequence is taken with long exposure (te = 0.4ms) and low framing (fs = 0.8kHz). The images are blurred with a Gaussian kernel and cropped along the ROI 1 (Fig. 7). The jet centerline is detected as the loci of points where the image is brighter. This is done by identifying the peak in the gray scale profile of each pixel column, previously smoothed with a Savitzky-Sgolay filter to increase differentiability. In the characterization of the impinged interface (2), both seeding injections in the gas flow are switched off and the video sequence is taken with long exposure (te = 0.4ms) and low framing (fs = 0.8kHz). As Rhodamine is diluted in the liquid, the laser induced fluorescence highly contrasts the liquid interface from the background. The liquid interface is detected using a convolution with a Sobel Kernel in an image crop ROI 2 (Fig. 7). In the characterization of the entrainment flow (3), both the seeding injections of the gas flow are switched on and the video sequence is taken with short exposure (te = 40µs) and high framing (fs = 4.6kHz). The velocity field is retrieved via Particle Image Velocimetry (PIV). Keeping ReM =4.5, Z/d=15, Rej =1800 for three cases with fp = 12, 14, 16Hz, Fig. 8 a) compares the horizontal displacement power spectra of a point located in the liquid interface, at the height of the centerline; Fig. 8 b) compares the vertical displacement power spectra of a point located in the jet centerline, between the nozzle and the wall. The liquid rib oscillates harmonically with the perturbation introduced, and an harmonic component appears in the jet response. However, the spectra of the jet is dominated by a frequency of the order of 60Hz, regardless of the perturbation introduced. Moreover, if the formation of the rib in the liquid Fig. 7: Example snapshot and film is avoided –by slightly increasing the stand off distance– no appreciable region of interest (ROI) for the jet oscillations occur. This suggests that there are two different destabilizing processing techniques mechanisms: one related to the asymmetry of the confinement, the other reimplemented. lated to the oscillation of such asymmetry. Fig. 8: a) Frequency content of the liquid rib produced by the impingement of the jet and pulsed by the perturbation introduced, b) frequency content of the corresponding jet response to the perturbation. Both mechanisms are characterized by recirculation structures produced on the side of the asymmetry: above the impingement in this simplified configuration (Fig. 7), below the impingement in the jet wiping configuration (Fig. 4). In both configurations, the jet periodically deflects towards the side where they occur. In conclusion, this work has rejected several hypotheses on the origin of undulation patterns in the final coating of jet wiped product. Moreover, it has shown that the unsteady confinement of the jet, due to the run back flow, can promote jet oscillation and therefore promote coating undulation. The dynamics of the two flows are linked by a hydrodynamic feedback which manifests through recirculation structures. Future work will assess the role of these structures in sustaining the jet oscillation. Acknowledgments M.A. Mendez is supported by a F.R.S.-FNRS FRIA grant and the project is funded by ArcelorMittal. References 1. Tuck, E.O, Continuous Coating with gravity and jet stripping, Phys. Fluids, 26 (9) 12, 2352–2358 (1983). 2. Hocking, G.C., Deformations during jet-stripping in the galvanizing process, J.Eng.Math, 70, 12, 297–306 (2011). 3. Yoon, H.G., Ahn, G.J., Kim, S.J., Chung, M.K., Aerodynamic Investigation about the Cause of Check-Mark Stain on the Galvanized Steel Surface, ISIJ International, 49 (11), 1755-1761 (2009). 4. Yoon,H.G., Chung,M.K., Development of Novel Air-knife System to Prevent Check-Mark Stain on Galvanized Strip Surface, ISIJ International, 50 (5), 752-759 (2010). 5. Peng,L., Chen,H., Vibration Analysis of Steel Strip in Continuous Hot-Dip Galvanizing Process, JAMP, 49 (11), 17551761 (2013). 6. Myrillas K., Rambaud P., Mataigne J-M., Gardin P., Vincent St., Buchlin J-M., Numerical modeling of gas-jet wiping process, Chem. Eng. Process., 68, 26–31 (2013).