降落液膜中PLIF,PTV 和红外测温检测方案(气体流量计)

智能文字提取功能测试中

ResearchGateSee discussions， stats， and author profiles for this publication at： https：//www.researchgate.net/publication/324594996 3rd Thermal and Fluids Engineering Conference (TFEC)ASTFEAmerican Societyof Thermal and Fluids EngineersMarch 4-7， 2018Fort Lauderdale， FL， USATFEC-2018-Xxxxx A combined experimental and computational study of the heat transfercharacteristics of falling liquid-films Conference Paper·March 2018DOI： 10.1615/TFEC2018.fmp.021505 CITATION1 READS240 5 authors， including： Alexandros CharogiannisImperial College London Fabian Denner Otto-von-Guericke-Universitat Magdeburg 36 PUBLICATIONS 167 CITATIONS 50 PUBLICATIONS 269 CITATIONS SEE PROFILE SEE PROFILE Berend van Wachem Christos N. Markides Otto-von-Guericke-Universitat Magdeburg Imperial College London 149 PUBLICATIONS 2，595 CITATIONS 266 PUBLICATIONS 2，279 CITATIONS SEE PROFILE SEE PROFILE Some of the authors of this publication are also working on these related projects： A COMBINED EXPERIMENTAL AND COMPUTATIONAL STUDY OF THEHEAT TRANSFER CHARACTERISTICS OF FALLING LIQUID-FILMS Alexandros Charogiannis， Fabian Denner， Berend G.M. van Wachem，Serafim Kaliadasis，Christos N. Markides，* Clean Energy Processes (CEP) Laboratory， Department of Chemical Engineering， Imperial College London， South Kensington Campus， London SW7 2AZ， United Kingdom Department of Chemical Engineering， Imperial College London， South Kensington Campus， London SW72AZ， United Kingdom Department of Mechanical Engineering， Imperial College London， South Kensington Campus， LondonSW7 2AZ， United Kingdom ABSTRACT An optical technique that combines planar laser-induced fluorescence (PLIF)， particle tracking velocimetry(PTV) and infrared thermography (IR) was applied for the investigation of the hydrodynamic and heat transfercharacteristics of harmonically-excited liquid films falling under the action of gravity over an inclined， elec-trically heated glass-substrate. PLIF was used to recover film-height data， PTV to recover two-dimensional(2D) velocity-field data， and IR to recover the temperature of the gas-liquid interface. The experiments werecomplemented by direct numerical simulations (DNSs) that provide additional information on the liquid vis-cosity， temperature and velocity distributions between the flow inlet and the downstream location where theoptical measurements were collected. By adoption of this synergistic approach we recover a wealth of infor-mation， including novel results on the spatiotemporal evolution of the interface topology， and the flow andtemperature fields underneath the wavy interface. Based on this data we also deduce local and instantaneousheat transfer coefficients (HTCs)， and focus our efforts towards the investigation of two HTC-enhancementmechanisms； the observation of“hot-spots”as precursors to the formation of thermal rivulets， which canresult in local enhancements in excess of 50%， and the presence of large velocity components in the cross-stream direction of the flow， which promote mixing and are shown to improve heat transfer by up to ≈ 7%compared to flow regions of the same height. KEY WORDS： Film flows， unsteady heat transfer， laser-induced fluorescence， particle velocimetry， infrared ther-mography， direct numerical simulations 1.INTRODUCTION Wave formation on liquid-film flows has been the subject of intense research for several decades worldwidewith experimental2]，numerical [and theoretical sstudies， both from the fundamental point of view， butalso for technological applications due to the heat and mass transfer characteristics of these flows. Applicationsinclude， amongst many others， various types of reactors， condensers， heat exchangers， wetted-wall absorbersand micro-cooling schemes. The advent of optical diagnostics in the fields of fluid mechanics and heat trans-fer has contributed significantly towards expanding our knowledge of a range of relevant flow-phenomena；yet， the availability of reliable experimental data from within falling liquid-films is relatively limited.This is primarily owing to the restricted fluid domains under observation and the intermittent nature of the movingand wavy interface which renders the application of even “conventional”optical techniques particularly chal-lenging. It is precisely the lack of experimental data relating to spatiotemporally resolved heat-transfer withinfalling-film flows that motivated the present study. The starting point is our earlier experimental， numericaland theoretical studies of isothermal falling-films6.0.1]，as well as the unsteady and conjugate heat transferin uniformly-heated falling films 1516]. (a) Water-ethanol flow (b) Water-glycerol flow Fig. 1 IR images of the film free-surface from flows comprising (a) 80% water and 20% ethanol by volume(Pr ≈ 5.5)， and (b) 45% water and 55% ethanol by volume (Pr~70)， with q~ 1.5 W cm-. Our experimental campaign comprises film flows with Reynolds numbers in the range Re =15 - 65， inletforcing-frequencies (i.e. wave frequencies) fw = 7 Hz， 12 Hz and 17 Hz， and heat-fluxes corresponding toq= 1.5 -3 W cm-2. Optical measurements are also conducted under isothermal film-flow conditions， atthe same flow Re and fw， in order to examine the impact of heating on the hydrodynamic characteristics ofthese films. The liquid we employ is a solution of glycerol (55% by volume) and water (45% by volume)，with a kinematic viscosity of v=9.2×10-6 m²s-1， a density of p = 1147 kg m-3， and a surface tensionof o=60×10-3Nm-1at 22 ℃. These properties were measured by collecting liquid samples throughpoutthe optical-measurement runs， while the liquid thermal conductivity，入=0.37 Wm-K-1， and thermaldiffusivity， K=1.2×10-7m²s-1， were estimated using the Fillipov equation in the first instance， and linearinterpolation between the values correpsonding to pure water and glycerol， resulting in a Prandtl number，Pr ~ 70. Finally， based on the liquid properties and the substrate inclination angle β=20 °， a liquid Kapitzanumber， Ka=330， is quoted. To investigate the hydrodynamic and heat transfer characteristics of harmonically-excited liquid films， wedeveloped an optical technique that combines planar laser-induced fluorescence (PLIF)， particle tracking ve-locimetry (PTV) and infrared thermography (IR). PLIF was used to recover space- and time-resolved film-height data， PTV to recover two-dimensional (2D) velocity-field data， and IR to recover the temperature of thegas-liquid interface. The experiments were complemented by direct numerical simulations (DNSs) that pro-vide additional information on the liquid viscosity， temperature and velocity distributions between the flowinlet and the downstream location where optical measurements were collected， and also allow us to expand theparameter space of our experiments. Before advancing to the main body of this research article， it is essentialto comment on the fluid selection. The use of a high glycerol-content aqueous solution has two noteworthyeffects on the heat transfer of thin liquid-films： First， it results in a low-conductivity (high-Pr) fluid whichsuppresses the formation of thermal rivulets (see， for example， Figure). The latter featured prominently inour previous study of the heat transfer characteristics of water-ethanol films (Pr ~ 5.5). Second， the liq-uid viscosity varies by approximately one order of magnitude across the liquid domain (see， for example，Cheng ⑧])， which is expected to significantly impact the film topology and flow field compared to an equiva-lent isothermal flow. Also of interest is the impact thereof on the variation of the liquid flow-rate with the filmheight， which for a periodic flow was shown to follow a simple linear relation (see， for example， Charogianniset al.C). This will be thoroughly examined in an upcoming publication. (a) Photograph of the opti-cal setup. (b) Schematic of the test sec-tion from top. (C) Schematic of the test sec-tion from the front. Fig. 2 (a) Photograph of the heated test-section and camera arrangement. (b) Schematic of the test-sectionshowing the relative orientation of the PLIF/PTV cameras and laser-sheet from above. (c) Schematic of thetest-section showing the relative orientation of the PLIF/PTV/IR cameras and laser-sheet from the front. 2. EXPERIMENTAL METHODOLOGY (a) PLIF frame of a solitary wave. (b) PIV velocity-vector field of a solitary wave. Fig. 3 Perspective-distortion corrected PLIF and PIV frames of a solitary wave from a flow with fw= 7 Hz，Re=55 and q=2.5 Wcm-2. Our flow facility comprises a closed flow-loop via which the liquid circulates， and a heated test-section overwhich liquid-film flows develop. The latter includes a flow-settling chamber which is installed at the flow inletin order to dispense the liquid evenly along the span of a glass substrate， which measures 300×300×1.1mmand is coated， on its top side (i.e. the one that is in contact with the liquid flow)， with an electrically-conductiveITO layer. The glass plate is held firmly on a hinged plastic sub-frame by a pair electrodes that extend along z [mm] Fig. 4 Instantaneous temperature distribution of the film free-surface of a flow with fw=7 Hz， Re = 28 andq=2.5W cm- its sides. The latter are connected to a 5.5 kW， DC power-supply that allows us to vary the heat input to theflow in the range intended for the experiments. The wave frequency is adjusted by means of an electronicallycontrolled throttling-valve that is installed upstream of the flow inlet. The liquid is seeded with 10-umparticles (glass hollow-spheres) and Rhodamine-B dye for conducting thePTV and PLIF measurements respectively. The dye-doped liquid is excited from below (i.e. the side that isnot in contact with the flow) using a frequency-doubled Nd：YAG laser (emission at 532 nm)， operated at100 Hz. A pair of CMOS cameras are installed underneath the test section in order to collect PLIF and PTVimages， while temperature measurements of the film free-surface are recovered using a FLIR X6540SC IRcamera which is set up atop the flow. The main optical components of our experimental rig are displayed inFigure 2(a)， while schematics of the test section and optics are provided in Figure2(b) and (c). To conduct film height and flow-velocity measurements simultaneously and at the same region of the flow， aswell as to correct for the refractive index mismatch between the employed liquid and surrounding air， the imag-ing planes of the PLIF and PTV cameras were mapped using a calibration target that was immersed inside theemployed water-glycerol solution. The spatial resolution of our PLIF/PV setup amounts to 28 m/pixel， whilethe downstream location where data were collected extends between 232 mm and 265 mm from the flow inlet.The raw PLIF images were corrected for optical distortions in La Vision Davis and then processed in MATLABusing an in-house developed algorithms in order to yield local and instantaneous film-height data， while theparticle-reflection images were used to generate 2-D velocity-vector distributions， following optical-distortioncorrections， by means of a four-pass cross-correlation algorithm available in LaVision Davis. Individual par-ticles were also tracked (PTV calculation) and used for further processing， including the calculation of localand instantaneous bulk-velocities and flow rates. Sample PLIF and PIV images are provided in Figure 3 for aflow with fw =7 Hz， Re =55 and q=2.5. The accuracy of the film-height， velocity and flow rate measurements， the latter obtained by combining the re-sults from the PLIF and PTV measurements， corresponds to ≈ 4%， based on a series of validation runs. Thesewere conducted in unperturbed， and therefore nearly flat， isothermal-films which allow for a priori knowl-edge of the mean film-height and bulk velocity. In the case of the IR measurements， perspective-distortioncorrections were applied once more， as the camera was set up at an angle to the substrate in order to pre-vent it from capturing an image of itself (due to the reflectivity of the liquid). Following this processing step，the radiation-intensity measurements were translated to 2-D temperature maps of the liquid surface using thecamera calibration-curve corresponding to the employed image-integration time. The accuracy of the IR mea-surement exceeded 1 °C based on quality-assurance experiments that were carried out using thermocouples.Fine K-type thermocouples were also employed in simultaneously recovering the solid-liquid interface tem-perature at the boundaries of the PLIF/PV imaging region. A sample interface-temperature measurement ispresented in Figure 4. also showing the location of the PLIF/PTV imaging region and thermocouples that areused to measure the liquid temperature at the wall. 3. DIRECT NUMERICAL SIMULATIONS METHODOLOGY Direct numerical simulations (DNS) of the entire two-phase flow system are conducted by resolving all rel-evant length and time scales. For incompressible flow of a Newtonian fluid the governing equations are thecontinuity， momentum and energy equations， which are solved using the finite-volume framework for inter-facial flows by Denner and van Wachem [C2]. The Volume-of-Fluid (VOF) method is adopted to capture theinterface between the interacting bulk phases [[4]. The interface is transported by the underlying flow using acompressive VOF method [[] and the fluid properties (density， viscosity， specific heat capacity and thermalconductivity) are defined based on the local volume fraction， as a linear average of the properties of both bulkphases. The density and specific heat capacity of each bulk phase are assumed to be constant. The temperature-dependent viscosity is obtained from the empirical model of Cheng ]， and the temperature-dependent ther-mal conductivity is computed based on the work of Bates 3]. Assuming surface tension is a volume forceacting at the interface， the surface force per unit volume is described by the Continuum Surface Force (CSF) Fig. 5 Sample film height， h， interface velocity， Uh， and interface temperature， Th， simulation-results for aflow with fw =17 Hz， Re=33 and q=3.0 W cm-2， presented as a function of the distance from the flowinlet， a， along the experimental and computational domain. model 图，which has previously been used frequently， and with great success， in the simulation of fallingliquid-films ]. Thermocapillary effects are neglected， since preliminary results showed they have a negligi-ble influence on the hydrodynamics and heat transfer of the considered falling-liquid films [13.1718]； themaximum temperature-difference observed in the simulations does not exceed 4 ℃， and the ensuing variationof the surface-tension coefficient due to thermocapillary effects is <0.5%(assuming a temperature-dependentsurface tension coefficient of or=8.5×10-5Nm-1K for the considered aqueous-glycerol mixture). Fig. 6 Experimentally recovered local and instantaneous HTCs， a， plotted against the respective film-heightdata， h， for flows flows with fw=7 Hz， q= 3.0 W cm-2and Re = 63， 45 and 25. The colour scalerepresents the probability density. 4. RESULTS AND DISCUSSION 4.1 Film Topology Sample film height， h， interface velocity， Uh， and interface temperature， Th， simulation-results are presented inFigure 5 for a flow with fw =17 Hz， Re =33 and q=3.0 W cm-2. In greater detail， the liquid-film height，velocity and temperature are tracked along the length of the entire experimental test-section (the location ofthe IR and PLIF/PTV imaging region is also shown for reference)， showing their evolution as more and moreheat is extracted with increasing distance from the flow inlet， c. Based on our numerical data， we observethat with increasing distance from the flow inlet， the mean (i.e. time-averaged) film-height falls as the liquidheats up and its viscosity drops， while the mean interface velocity， as well as the bulk velocity， both increasein order to preserve continuity. The gas-liquid interface temperature only starts to increase at o ≈ 150 mm，owing to the low conductivity of the employed water-glycerol solution， while any heat extracted from thewall penetrates through the film and all the way to the gas-liquid interface along the thinner film regions first.Finally， for the wave frequency we set in this case(i.e. fw=17 Hz)， the wave topology evolves throughout thelength the computational domain due to the spatiotemporal variation of the liquid viscosity， with the numberof capillary waves increasing from 0 over the range c≈ 0- 150 mm， to 1 over the range ≈175-225 mm，and finally to 2 over the range a≈ 225-300 mm.This continuous evolution of the interface topology， whichwas also confirmed in our experiments despite the limited span of our PLIF/PTV imaging domain， was lesspronounced at lower wave frequencies (i.e. fw = 7 Hz and fw= 12 Hz). In addition to the variation of the flow and liquid properties as a function of the distance from the flowinlet， when observed locally (i.e. at the PLIF/PTV imaging region)， the former are strongly dependent on theimposed pulsation-frequency， the heat flux that is applied on the wall side， and the flow Re. Specifically， withincreasing wave frequency from fw = 7 Hz to fw = 17 Hz and for a constant Re and q， the wave crest-heightfalls and the trough height increases， as do the local bulk and interface velocities. With increasing heat fluxfrom q=0 W cm-2 to q=3.0 W cm-2 and for a constant fw and Re， the mean and maximum film-heights(i.e. the crest heights) fall， while the bulk and interface velocities increase， once again due to the reduction inthe liquid viscosity. In contrast， the local flow-rates remain constant along the flow regions that retain theirtopology (for example， the main wave humps). Finally， with increasing flow Re， the mean film-heights andbulk velocities increase， resulting in a pronounced reduction in the observed gas-liquid interface temperatures. 4.2 Heat Transfer Characterization Earlier in the Introduction we noted that the high Pr flows that we investigate in this synergistic experimen-tal/computational campaign suppress the formation and growth of thermal rivulets. Thus， the spatiotemporalvariation of the heat transfer coefficient (HTC) is primarily governed by the periodic variation of the filmheight and bulk velocity， which drives the periodic variation of the gas-liquid and solid-liquid interface tem-peratures. The latter is far less prominent than the former， to the extent that it cannot be retrieved using theemployed instrumentation (i.e. the thermocouples that are installed on the heated wall， immediately upstreamand downstream of the PLIF/PTV imaging domain). The wave topology is governed by the mechanism ofsurface-wave generation and amplification， as well as the reduction of the liquid viscosity along and acrossthe examined films. The ensuing relation between any large-amplitude fluctuations of the HTC， a， and thefilm height， h， is demonstrated in Figure 6 which shows bivariate distributions of the local and instantaneousHTCs and film-heights for flows with fw = 7 Hz，q=3.0 and Re =63， 45 and 25 (the colour scale repre-sents the probability density). These plots were generated from HTC and film-height time-series which were，in turn， generated by averaging film-height and temperature measurements locally，over a small region of theflow (~ 1.4 mm)， on aper-image basis. The HTC time-series were generated based on the gas-liquid interfacetemperature data， Th， that were recovered using IR thermography， the solid-liquid interface temperature data，Ts， that were recovered using thermocouples， and the mean heat-fluxes， q， that were applied to the liquid films When the film-height (or equivalently the bulk velocity or flow rate) falls， the temperature difference acrossthe liquid domain decreases and the HTC increases. Like we noted earlier， the flow away from the wall canbe sufficiently decoupled from the flow near the wall， for example near the flow inlet， to the extent that thegas-liquid interface temperature is insensitive to any film-height fluctuations. This is also the case at highRe (see， for example， Figure b(a)， (b)). At low Re， however， the HTC is strongly coupled to h， particularlyalong the thinner film-regions， as any heat extracted from the wall penetrates more easily through the filmand all the way to the free surface (see， for example， see， for example， Figure b(c)). Films excited at 7 Hzexhibit extended thin-film regions and locally higher HTCs compared to films excited at higher frequencies，for example fw= 12 Hz and 17 Hz， as well as larger solitary waves， and thus locally lower HTCs near thewave crests. Thus， such a topology contributes a broader HTC range. 4.3 Heat Transfer Enhancement Based on the collected experimental data and the associated numerical simulations， we have identified andare hereby reporting on two local heat-transfer enhancement mechanisms； one stemming from the locallyincreased free-surface temperatures ahead of the main wave humps compared to film regions of the sameheight from behind the waves， and one that relates to the observation of highly localised“hot spots” along thethinner film regions， at low Re and high q. The former can be clearly observed in the probability distributionswe present in Figure B. and correspond to the low probability data-points found between the solitary-wavecrest and trough， above the main body of data. As evidenced from these plots， the HTC increases locally by~5-7%. Using DNSs， weoverlap cross-stream velocity distributions from flows with q=3.0 and Re = 33and fw = 17 Hz， 22 and 30 Hz， with isotherms drawn along the waves over the range x = 220- 280 mm(see， Figure7). The cross-stream velocities peak ahead of the main wave humps and then drop to negativevalues， reaching a minimum behind the capillary waves. Thus， hot liquid from near the wall is pushed towardsIm free-surface， promoting mixing. In Figure 8 we wtea ktea kae cal ocsleors elro olko oakt at this behaviour by plotting thecross-stream velocity， Uy，h， against the interface temperature， Th， for the q=3.0， Re = 33 and fw = 17 Hzand 30 Hz flows over one wavelength (i.e. for x≈ 237- 264 and o ≈ 240- 260 mm， respectively). In x[mm] Fig. 7 Instantaneous cross-stream velocity， Uy， distributions overlapping isotherms drawn along the waves offlows with fw= 30 Hz (top)， fw = 22 Hz (middle) and fw = 17 Hz (bottom)， and Re =28 and q=3.0W cm-2 over the range x=220 - 280 mm. These results were obtained by DNS. both cases as Uy，h increases from Uy，h=0 to the peak value between the wave crests and troughs， Th remainsnear-constant (and equal to the peak interface-temperature) and subsequently falls ahead of the wave troughs.The additional“branches”that are observed in the case of the fw = 17 Hz flow originate from the capillary-wave region. Both the Th and Uy，h fluctuations are smaller for the fw =30 Hz flow which suggests that anyheat-transfer gain is lower. The local enhancement of the HTC and its dependence on the local value of Uyhwas also confirmed using the experimental data. (a) (b) Fig. 8 Cross-stream velocities at the gas-liquid interface， Uy，h， plotted against the interface temperature， Th，for flows with q=3.0 W cm-2， Re =33 and (a)fw = 17 Hz and (b) fw = 30 Hz flows， over one spatialperiod. These results were obtained by DNS. With regard to the observation of hot spots in the IR images， these interfacial features originate from hotpackets of fluid that detach from the hot substrate and rise to the liquid surface while being convected by themean flow. Essentially a precursor to the formation of thermal rivulets (see， for example， Refs.[15，16])， theyare typically encountered in the substrate-film regions and are randomly scattered across the field of viewof the IR camera. The interface-temperature of these features can exceed that of the surrounding fluid bymore than 10℃， resulting in HTC enhancements in excess of 50%. In Figureg(a) we present an interface-temperature map showing a pair of hot-spots from a flow with q=3.0， Re = 25 and fw = 7Hz， while inFigure9(b) we overlap the coordinates of all the hot-spots we have identified over the entire measurement runalong the liquid domain (in the axial direction of the flow) where we conduct the PLIF/PTV measurements，with the time-averaged temperature field. In Figure 10(a)， we plot the maximum local interface-temperature(i.e. the maximum temperature of the hot-spots) against the number of hot-spots we have identified for thesame flow/heating condition， as a function of the spatial domain over which we average our temperature data(in pixels). With increasing size of the averaging domain， the peak of the distribution moves towards lowertemperatures and the number of occurrences of high-temperature spots falls； yet the majority of the plot-ted temperature-data bear higher values compared to the peak interface-temperature of the film which can befound at the wave-trough and which for the same flow/heating conditions amounts to ≈ 30.5C. With increas-ing flow Re， both the number and the peak temperature of the hot-spots falls significantly (see， Figure 10(b))，with almost no hot-spots observed already from Re =54， owing to the increased flow velocities. 5. CONCLUSIONS We have presented flow and heat transfer results from the application of a combined optical technique toharmonically-excited liquid films falling under the action of gravity over an inclined， electrically heated glass-substrate. Planar laser-induced fluorescence (PLIF) was used to recover film-height data， particle tracking ve-locimetry (PTV) to recover 2D velocity-field data， and infrared thermography (IR)to recover the temperatureof the gas-liquid interface. The experiments were complemented by DNSs that provide additional informa-tion on the liquid viscosity， temperature and velocity distributions between the flow inlet and the downstreamlocation where the optical measurements were collected. By adoption of this synergistic approach we recover (a) (b) Fig. 9 Instantaneous， (a)， and time-averaged， (b)， interface-temperature maps showing the coordinates wherehot-spots were identified in a flow with q=3.0 W cm-2， Re= 25 and fw =7 Hz. (a) (b) Fig. 10 (a) Hot-spot temperatures plotted against the number of hot spots for a flow with fw=7 Hz， Re = 25and q=3.0 Wcm-. (b) Maximum hot-spot temperatures plotted against the number of hot spots for flowswith fw= 7Hz， q= 3.0 W cm-2 and Re=25-54. a wealth of information， including novel results on the spatiotemporal evolution of the interface topology， andthe flow and temperature fields underneath the wavy interface. In greater detail， we note that the films becomethinner and accelerate with increasing distance from the flow inlet， while any heat extracted from the wallpenetrates more easily through the film and all the way to the gas-liquid interface at lower Re and higher q.Using the same data we also calculate local and instantaneous heat transfer coefficients (HTCs)， and focusour investigation on two HTC-enhancement mechanisms； the observation of hot spots as precursors to theformation of thermal rivulets， which can result in local enhancements in excess of 50%， and the presence oflarge velocity components in the cross-stream direction of the flow， which promote mixing and are shown toimprove heat transfer by up to≈ 7% compared to flow regions of the same height. ACKNOWLEDGMENTS This work was supported by the UK Engineering and Physical Sciences Research Council (EPSRC) (grantnumbers EP/K008595/1，EP/L020564/1 and EP/M021556/1). NOMENCLATURE Heat transfer coefficient (Wm-2K-1) h Film height (m) Inclination angle (°) Kapitza number (-) Thermal diffusivity (m²s-1) Prandtl number (-) Thermal conductivity (Wm-1K-1) q Heat flux (Wm-2) Kinematic viscosity (m²s-1) Reynolds number (-) Density (kgm-3) Temperature (C) Surface tension (Nm-1) Flow velocity (ms-) Wave frequency ( Hz) oc Axial distance (m) REFERENCES [1]AAlbert， C.，Raach， H.， and Bothe， D.， “Influence of surface tension models on the hydrodynamics of wavy laminar falling filmsin volume of fluid-simulations，” Int. J. Multiphase Flow， 43， pp. 66-71，(2012). [2]AAlekseenko， S. V.，Nakoryakov， V.E.， and Pokusaev， B. G.， “Wave formation on a vertical falling liquid film，"AIChE J.， 31(9)，pp. 1446-1460，(1985). [3]Bates，O.，“Binary mixtures of water and glycerol- thermal conductivity of liquids，” Ind. Eng. Chem. Chem.， 28， pp. 494-498，(1936). [4]Brackbill， J.， K. D. and Zemach， C.，“Continuum method for modeling surface tension，”J. Comput. Phys.， 39，pp. 335-354，(1992). [5]CChakraborty， S.， Nguyen， P.-K.，Ruyer-Quil， C.， and Bontozoglou， V.， “Extreme solitary waves on falling liquid films，"J. FluidMech.，745， pp. 564-591，(2014). [6] Charogiannis， A.， An， J. S.， and Markides，C. N.， “A simultaneous planar laser-induced fluorescence，particle image velocimetryand particle tracking velocimetry technique for the investigation of thin liquid-film flows，” Exp. Therm. Fluid Sci.， 68， pp.516-536，(2015). ( [7] Charogiannis， A . ， Denner，F.， van Wachem， B.， Kal l iadasis， S.， and Ma r kides， C.，“Detailed hydrodynamic characterization ofharmonically excited falling-film flows： A combined experimental a n d computational st u dy，”Phys. Rev. Fluids，2， pp . 014002- 1-37，(2016). ) ( .Cheng， N.-S.， “Formula f or the viscosity of a glycerol-wate r mixture， I nd. Eng. Chem. Res.， 47， pp. 3285-3288，(2008). ) ( Demekhin， E.， Kalaidin，E.， Ka l liadasis， S.， a n d Vlaskin， C.，“Three-dimensional localized coherent struc t ures of surface turbu-lence. i. scenarios of two-dimensional-three-dimensional transition，"Phys. Fluids， 19， pp.114 1 03，(2007 ) . ) ( [10] LDenner， F， Pradas， M . ， C h arogiannis， A. ， M ar kides， C. N .， van Wachem， B. G. M.， an d Kalliad a sis， S.，“Self-si m ilarity ofsolitary waves on inertia-dominated falling liquid films，” Phys. Rev. E， 93， pp.033121-1-14，(2016). ) ( [11] D [ enner， F . a nd van Wachem， B.，“Co m pressive vof m e tho d with s k ewness correction to capt u re sharp i n terfaces on arbitra r y meshes，”J. Comput. P hys.，279， pp.1 2 7-144， (2014). ) ( [12] Denner， F. and v an W achem， B. ， “ Fu lly-coupled bal a nced-force vof framework for arbitrary meshes with leas t -squares curvatureevaluation from v olume fractions，Numer. Heat Transfer Pa r t B， 65， pp. 2 1 8-255，(20 1 4). ) ( [13] L Deshpande， S.， A numolu， L . ， and Trujillo， M.， “ Evaluating the performance of the two-phase flow solver interfoam，” C omput Sci. Discovery，5，pp. 014016，(2012). ) ( [14] Hirt， C. and Nichols， B.，“Volume of fluid (vof) method for the dynamics of free boundaries，”J. Comput. Phys.， 39， pp. 201-225， (1981). ) ( [15] Markides， C. N . ， M a thie， R.， and Charogiannis， A.， “An experimental characterization of spatiotemporally resolved heat transferin thin liquid-film flows falling over a n inclined heated foil，”Int. J. Heat Mass Tran.， 9 3， pp. 8 72-888，(2015). ) ( [16] N Mathie， R.， Nakamura， H.， a nd Markides， C. N.，“Heat transfer augmentation in unstea d y conjugate thermal systems - Part II： Applications，”Int. J. Heat Mass Tran.， 56， pp. 819-833， ( 2013). ) ( [17] Nabil， M. and A.S.， R.，“A computational study on the effects of surface tension and prandtl number on laminar-wavy falling-filmcondensation，J. Heat Transfer， 139， pp. 121501-1-1 1，( 2016). ) ( [18] Raeini， A.， Blunt， M.， and Bijeljic ， B.，“Modelling two-phase flow in p orous media at the pore scale u sing t he volume-of-fluidmethod，”J. Comput. Phys.， 231， pp. 5653- 5 668， (2012). ) All content following this page was uploaded by Alexandros Charogiannis on April The user has requested enhancement of the downloaded file. *Corresponding Christos N. Markides： c.markides@imperial.ac.uk       An optical technique that combines planar laser-induced fluorescence (PLIF), particle tracking velocimetry (PTV) and infrared thermography (IR) was applied for the investigation of the hydrodynamic and heat transfer characteristics of harmonically-excited liquid films falling under the action of gravity over an inclined, electrically heated glass-substrate. PLIF was used to recover film-height data, PTV to recover two-dimensional (2D) velocity-field data, and IR to recover the temperature of the gas-liquid interface. The experiments were complemented by direct numerical simulations (DNSs) that provide additional information on the liquid viscosity, temperature and velocity distributions between the flow inlet and the downstream location where the optical measurements were collected. By adoption of this synergistic approach we recover a wealth of information, including novel results on the spatiotemporal evolution of the interface topology, and the flow and temperature fields underneath the wavy interface. Based on this data we also deduce local and instantaneousheat transfer coefficients (HTCs), and focus our efforts towards the investigation of two HTC-enhancement mechanisms; the observation of “hot-spots” as precursors to the formation of thermal rivulets, which can result in local enhancements in excess of 50%, and the presence of large velocity components in the crossstream direction of the flow, which promote mixing and are shown to improve heat transfer by up to  7% compared to flow regions of the same height.