Effect of Wind Stress on the Dynamics and Stability of Nonisothermal Power-Law Film down an Inclined Plane

Abstract

Dynamics and stability of a nonisothermal power-law liquid film down an inclined plane is considered in the presence of interfacial shear. Linear stability characteristics of the power-law liquid film using normal mode approach reveal that isothermal and evaporating films are unstable for any value of power-law index while there exists a critical value of power-law index for the case of condensate film above which condensate film ow system is always stable. This critical value of power-law index increases with the increase in shear stress at the interface. Weakly nonlinear stability analysis using method of multiple scales divulges the existence of zones due to supercritical stability and subcritical instability. The nonlinear evolution equation is solved numerically in a periodic domain. The results reveal that (1) for an isothermal dilatant (pseudoplastic) liquids, the maximum wave amplitude is always smaller (larger) than that for a Newtonian liquid and the amplitude of permanent wave increases with the increase in interfacial shear; (2) condensation of pseudoplastic film happens for the earlier instant of time when the phase change parameter increases and the effect of interfacial shear makes the film more corrugated; (3) dilatant (pseudoplastic) evaporating liquid film attains rupture faster (slower) than that of Newtonian liquid film, and the interfacial shear does not influence the time at which rupture occurs.

1. Introduction

Gravity-driven flow of a thin film down a vertical or an inclined plane has attracted much attention due to its importance in many industrial applications such as film coating and interface heat and mass transfer processes in chemical technology and energetics.Yih [1] and Benjamin [2] first studied the linear theory for the isothermal falling film. Yih [3] formulated the problem in terms of long-wave asymptotics and determined the critical Reynolds number above which the instability would occur. Benney [4] extended the theory to nonlinear regime by deriving a nonlinear evolution equation. There have been number of extensions of this work as discussed by Chang and Demekhin [5] in their monograph.

Although the theory of laminar film condensation flow due to gravity has been analyzed by Nusselt [6], the stability analysis of condensate or evaporating liquid film down a vertical or an inclined plane has been studied only after 1970s [7–9]. Burelbach et al. [10] and Bankoff [11] formulated a one-sided model and studied nonlinear stability and breakdown of evaporating/condensing horizontal static liquid film. Long-wave instabilities of heated falling films have been investigated by Bankoff [12] and Joo et al. [13]. Effect of interfacial phase change on the linear stability of thin films has been considered by Spindler et al. [14], Ünsal and Thomas [15], and Spindler [16]. They have showed that vapor recoil effects are destabilizing for a volatile liquid. Using perturbation methods, Ünsal and Thomas [17] have investigated the nonlinear stability of vertical condensate film flow. Hwang and Weng [18] have examined the finite-amplitude stability analysis of liquid film down a vertical plane with and without interfacial phase change and have shown that both supercritical stability and subcritical instability are possible for condensate film flow system.

These results for a Newtonian film flows (isothermal, condensate, evaporating) cannot completely describe the dynamics of non-Newtonian film flows. For example, mud flows are pseudoplastic, saccharified honey is known to be a shear thickening fluid, whereas polymers exhibit generally a large elastic component. Non-Newtonian fluids generally exhibit a nonlinear relationship between shear stress and shear rate. These flows may be classified as inelastic and viscoelastic. The inelastic fluids may be subdivided as time-dependent fluids (thixotropic and rheopectic) and time-independent fluids (pseudoplastic, dilatant, Bingham plastic, pseudoplastic with yield stress). Apart from investigation on flow characteristics and stability analysis of viscoelastic fluid films along a vertical or on inclined plane [19–27], inelastic time-independent non-Newtonian fluids have received the greatest attention from rheologists [28–33] which has resulted in the development of a number of equations or models proposed to represent their flow behavior. The Ostwald de Waele power-law model represents several inelastic time-independent non-Newtonian fluids of practical interest and therefore has been used in this paper.

The investigations on the stability characteristics of Newtonian fluid film down an inclined/a vertical plane show that in the linear theory, the film flow system is unstable for any Reynolds number. However, the finite-amplitude stability analysis of liquid films down a vertical wall by Hwang and Weng [18] with interfacial phase change reveals that the isothermal and evaporating Newtonian films are unstable for any Reynolds number while there exists a finite critical Reynolds number for the case of condensate film below which condensate Newtonian film flow system is always stable. This shows that the effect of mass transfer at the interface of a Newtonian fluid film strongly modifies the stability characteristics of the film flow when the phase change is considered. Usha and Uma [34] have extended their study to condensate/evaporating power-law liquid film down an inclined plane with or without interfacial phase change to show that isothermal and evaporating films are unstable for any value of power-law index “𝑛” while there exists a critical value of power-law index “𝑛” for the case of condensate film above which condensate film flow system is always stable. van der Walls interactions are taken into account by Gorla [35] to study the rupture of thin power-law liquid film on a cylinder. He has shown that rupture time for the dilatant fluids are higher than that of Newtonian and pseudoplastic fluids. This shows that a detailed numerical investigation on nonlinear evolution of nonisothermal power-law liquid film down an inclined plane is required to shed more light on understanding the effect of power-law index “𝑛” on different kinds of permanent wave shapes (isothermal film), film condensation (condensate film), and film rupture (evaporating film).

Recently, Pascal and D’Alessio [36] have studied the generation and structure of roll waves on the surface of a isothermal power-law liquid film down an inclined plane. To have a more realistic description of flows taking place in an environment, they have considered the effect of wind stress acting on the surface of a power-law film down an incline. Effect of interfacial-induced shear on thin films has been first attempted by Sheintuch and Dukler [37]. Later, Pascal [38] studied the effect of superficial shear stress in connection with the study of wind-aided spreading of oil on the sea.

The chief motivation of the present study is to investigate the dynamics and stability of more realistic flows taking place in an environment. In this case, it becomes important to include the effects of superficial shear stress and phase change at the interface of a fluid film flowing down an inclined plane. The Ostwald de Waele power-law model is considered to represent the non-Newtonian fluid.

In this paper, dynamics and stability of condensate/evaporating power-law liquid film flowing down an inclined plane with the effect of wind stress and phase change at the interface are considered. Linear stability analysis using normal mode approach reveals that cocurrent superficial wind stress at the interface destabilizes the film flow system. Weakly nonlinear stability analysis using method of multiple scales divulges the existence of zones due to supercritical stability and subcritical instability. The nonlinear evolution equation describing the shape of the free surface has been solved numerically in a periodic domain using Lee’s three-time level method. Applications of the results indicate that the amplitude of permanent wave increases with the increase in prescribed shear stress at the interface for isothermal power-law film and pseudoplastic liquid film has the largest amplitude compared to Newtonian and dilatant film. Also, evaporating liquid film attains rupture faster than that of Newtonian and condensate liquid film, and the prescribed shear stress at the interface does not influence the time at which rupture occurs.

2. Mathematical Formulation

A thin power-law liquid film flowing down an inclined plane 𝑦=0 with the effect of prescribed superficial shear stress and phase change acting on the surface of a fluid layer 𝑦=ℎ(𝑥,𝑡) (Figure 1) is considered.

Figure 1 

Schematic representation of a thin film down an inclined plane.

The governing equations [18, 24, 34] are the two-dimensional mass, momentum, and energy balance equations for the power-law model given by𝜕𝑢𝜕𝑥+𝜕𝑣𝜌𝜕𝜕𝑦=0,𝑢𝜕𝑡+𝑢𝜕𝑢𝜕𝑥+𝑣𝜕𝑢𝜕𝑦=−𝜕𝑝𝜕𝑥+𝜌𝑔sin𝜃(1−𝛾)+𝜕𝜏_𝑥𝑥𝜕𝑥+𝜕𝜏_𝑥𝑦𝜌𝜕𝜕𝑦,𝑣𝜕𝑡+𝑢𝜕𝑣𝜕𝑥+𝑣𝜕𝑣𝜕𝑦=−𝜕𝑝𝜕𝑦−𝜌𝑔cos𝜃(1−𝛾)+𝜕𝜏_𝑦𝑥𝜕𝑥+𝜕𝜏_𝑦𝑦𝜕𝜕𝑦,𝑇𝜕𝑡+𝑢𝜕𝑇𝜕𝑥+𝑣𝜕𝑇𝜕𝑦=𝜌𝐾𝑐_𝑝𝜕²𝑇𝜕𝑥²+𝜕²𝑇𝜕𝑦²,(2.1) where 𝐾 is the thermal conductivity, 𝜌 is the density, 𝑐_𝑝 is the liquid specific heat, 𝑔 is the gravity, 𝛾 is the ratio of vapor density to liquid density, and 𝑇 is the temperature.

The boundary conditions at the wall are the no-slip condition of velocity and a constant wall temperature (𝑇_𝑤) given by𝑢=0,𝑣=0,𝑇=𝑇_𝑤on𝑦=0.(2.2) The effect of wind stress is implemented into the model by prescribing the shear stress 𝜏_𝑤 at the free surface of the fluid layer [36]. The boundary conditions at the liquid-vapor interface [17, 18, 34, 36] are the balance of normal and tangential stresses, the relation of interfacial energy balances, and the equality of liquid and saturated vapor temperatures (𝑇_𝑠) and are given by𝑝+2𝜕ℎ𝜕𝑥𝜏_𝑥𝑦−𝜕ℎ𝜕𝑥²𝜏_𝑥𝑥−𝜏_𝑦𝑦1+𝜕ℎ𝜕𝑥²⁻¹+𝐾²ℎ^−2𝑓𝑔(𝛾−1)𝜌𝛾𝜕𝑇𝜕𝑦−𝜕ℎ𝜕𝑥𝜕𝑇𝜕𝑥²×1+𝜕ℎ𝜕𝑥²⁻¹+𝜎𝜕²ℎ𝜕𝑥²1+𝜕ℎ𝜕𝑥²^−3/2𝜕=𝑝_𝑎on𝑦=ℎ,ℎ𝜕𝑥𝜏_𝑦𝑦−𝜏_𝑥𝑥+1−𝜕ℎ𝜕𝑥²𝐾𝜕𝜏_𝑥𝑦=𝜏_𝑤on𝑦=ℎ,𝑇𝜕𝑦−𝜕ℎ𝜕𝑥𝜕𝑇𝜕𝑥−𝜌ℎ^𝑓𝑔𝜕ℎ𝜕𝑡+𝑢𝜕ℎ𝜕𝑥−𝑣=0on𝑦=ℎ,𝑇=𝑇_𝑠on𝑦=ℎ,(2.3) where𝜏_𝑥𝑥=2𝜇_𝑛𝜕𝑢𝜕𝑥𝜕𝑢𝜕𝑦^𝑛−1,𝜏_𝑥𝑦=𝜏_𝑦𝑥=𝜇_𝑛𝜕𝑢𝜕𝑦^𝑛,𝜏_𝑦𝑦=−2𝜇_𝑛𝜕𝑢𝜕𝑥𝜕𝑢𝜕𝑦^𝑛−1,(2.4)𝜇_𝑛 is the consistency coefficient, and 𝑛 is the flow index. When the power-law exponent 𝑛 is equal to 1, then the model describes the Newtonian fluid; If 𝑛<1, the fluid is said to be pseudoplastic or shear thinning, and, if 𝑛>1, the fluid is called dilatant or shear thickening. Using the dimensionless quantities defined by𝑢=𝑢𝑢₀,𝑣=𝑣𝜅𝑢₀(velocitiesinthe𝑥and𝑦directionsresp.),𝑥=𝜅𝑥ℎ₀,𝑦=𝑦ℎ₀𝑡=𝜅,𝑢₀𝑡ℎ₀(time),𝐻=ℎℎ₀(filmthickness),Fr=𝑢²⁰𝑔(1−𝛾)ℎ₀(Froudenumber),𝑢₀=𝑛𝑛+1𝜌𝑔(1−𝜇_𝑛𝛾)sin𝜃^1/𝑛ℎ^{0(𝑛+1)/𝑛}(referencevelocity),𝜅=2𝜋ℎ₀𝑝=𝜆(wavenumber),𝑝−𝑝_𝑎𝜌𝑢²⁰(pressure),Θ=𝑇−𝑇_𝑤𝑇_𝑠−𝑇_𝑤(temperature),Re_𝑛=𝜌𝑢^02−𝑛ℎ^𝑛0𝜇_𝑛Pr_𝑛(Reynoldsnumber),=𝜇^𝑛𝑢^0𝑛−1ℎ^01−𝑛𝑐_𝑝𝐾(Prandtlnumber),Pe^𝑛𝜉=𝑐_𝑝=Pr_𝑛⋅Re_𝑛(Pecletnumber),Δ𝑇=𝑇_𝑠−𝑇_𝑤,Δ𝑇ℎ_𝑓𝑔ℎ_𝑓𝑔−latentheat,Fi_𝑛=𝜎𝜌𝑢²⁰ℎ₀=We_𝑛Re^{𝑛−(3𝑛+2)/(𝑛+2)}1sin𝜃^{(3𝑛−2)/(𝑛+2)}𝑁𝑑=𝜉₂,𝛾Pr^2𝑛,𝜏=𝜏_𝑤We_𝑛𝜌𝑔ℎ₀sin𝜃(Shearstressattheinterface),=𝜎𝜌⁻¹𝑛+𝑛1^{𝑛(3𝑛−2)}[(1−𝛾)𝑔]^{−(3𝑛−2)}𝜈⁻⁴^1/(𝑛+2)(Webernumber),(2.5) the nondimensional governing equations and the boundary conditions are obtained as𝜕𝑢𝜅𝜕𝑢𝜕𝑥+𝜕𝜕𝑣𝑦=0,(2.6)𝜕𝑡+𝑢𝜕𝜕𝑥𝑢+𝑣𝜕𝜕𝑦𝑢+𝜕𝜕𝑥𝑝=siFnr𝜃+R𝜅e²_𝑛𝜕𝜏_𝑥𝑥𝜕𝑥+R1e_𝑛𝜕𝜏_𝑥𝑦𝜅𝜕𝑦,(2.7)₂𝜕𝜕𝑣𝑡+𝜕𝑢𝜕𝜕𝑥𝑣+𝑣𝜕𝜕𝑣𝑦=−𝜕𝜕𝑦𝑝−coFsr𝜃+R𝜅e_𝑛𝜕𝜕𝜏𝑥^𝑦𝑥+𝜕𝜕𝜏𝑦^𝑦𝑦,(2.8)₂Θ𝜕𝑦²=𝜅Pe_𝑛𝜕𝜕Θ𝑡+𝑢𝜕𝜕Θ𝑥+𝑣𝜕𝜕Θ𝑦−𝜅²𝜕²Θ𝜕𝑥²𝑝+1𝑢=0,𝑣=0,Θ=0on𝑦=0,,(2(.21.09))Re_𝑛2𝜅𝜕𝜕𝐻𝑥𝜏^𝑥𝑦−𝜅𝜏_𝑦𝑦−𝜅³𝜕𝐻𝜕𝑥²𝜏_𝑥𝑥1+𝜅²𝜕𝜕𝐻𝑥²⁻¹+𝜅²Fi_𝑛𝜕²𝐻𝜕𝑥²×1+𝜅²𝜕𝐻𝜕𝑥²^−3/2+𝑁𝑑Re^𝑛−2(𝛾−1)𝜕Θ𝜕𝑦−𝜅²𝜕𝐻𝜕𝑥𝜕𝜕Θ𝑥²1+𝜅²𝜕𝜕𝐻𝑥²⁻¹=0on𝑦𝜅=𝐻,(2.11)₂𝜕𝐻𝜕𝑥𝜏^𝑦𝑦−𝜏_𝑥𝑥+1−𝜅²𝜕𝐻𝜕𝑥²𝜏^𝑥𝑦=Re^𝑛sin𝜃𝜉𝜕ΘFr𝜏on𝑦=𝐻,(2.12)𝜕𝑦−𝜅²𝜕𝐻𝜕𝑥𝜕𝜕Θ𝑥−Θ𝜅P=e^𝑛1𝜕𝜕o𝐻n𝑡+𝑦=𝑢𝜕𝜕𝐻𝐻𝑥.−𝑣=0on𝑦=𝐻,((22..1134)) In (2.13), 𝜉>0 corresponds to condensate film flow (saturated vapor temperature is more than the wall temperature; (𝑇_𝑠>𝑇_𝑤)), 𝜉<0 corresponds to evaporating film flow (saturated vapor temperature is less than the wall temperature; (𝑇_𝑠<𝑇_𝑤)), and 𝜉=0 corresponds to isothermal film (saturated vapor temperature is equal to the wall temperature; (𝑇_𝑠=𝑇_𝑤)).

It is to be noted that, when 𝑛=1,𝜏=0,𝜃=90^∘, the above equations and boundary conditions reduce to evaporating or condensating Newtonian flow down a vertical wall investigated by Hwang and Weng [18] and, for 𝜏=0, they reduce to the equations obtained for nonisothermal power-law fluid film flow investigated by Usha and Uma [34]. Since the long wavelength modes are the most unstable ones for the film flow, the physical quantities 𝑢,𝑣,𝑝, and Θ are expanded in powers of small wave number 𝜅. Substituting these in (2.6)–(2.14) and collecting the coefficients of like powers of 𝜅, the zeroth- and the first-order equations are obtained. Noting that We_𝑛 is large in practical applications, 𝜅²We_𝑛 is taken to be of order one. Also, since the effect of 𝑁𝑑 has been found to be negligible on stability [15], 𝑁𝑑Re^𝑛−2 is taken to be of order 𝜅². Further, in the analysis, Re_𝑛≃𝑂(1) and Pe_𝑛≃𝑂(1). The first-order solutions, 𝑢₁,𝑣₁, and Θ₁, will have the time derivative 𝐻_𝑡 in the right hand side, which is replaced by the space derivative from the leading order representation [18] given by𝐻_𝑡=𝜉𝜅Pe_𝑛Θ_0𝑦−𝑢₀𝐻_𝑥+𝑣₀.(2.15) The zeroth- and the first-order (after replacing all the time derivatives with (2.15)) solutions are presented in the appendix.

The solutions (A.1) and (A.2) are substituted into the expanded kinematic boundary condition (2.13) yields:𝐻_𝑡=𝜉𝜅Pe_𝑛Θ_0𝑦+𝜅Θ_1𝑦−𝜅²𝐻_𝑥Θ_0𝑥+𝜅Θ_1𝑥−𝑢₀+𝜅𝑢₁𝐻_𝑥−𝑣₀−𝜅𝑣₁.(2.16) The simplification of (2.16) now gives the generalized kinematic equation as𝐻_𝑡+𝑋(𝐻)+𝐴(𝐻)𝐻_𝑥+𝐵(𝐻)𝐻_𝑥𝑥+𝐶(𝐻)𝐻_{𝑥𝑥𝑥𝑥}+𝐷(𝐻)𝐻^2𝑥+𝐸(𝐻)𝐻_𝑥𝐻_𝑥𝑥𝑥+𝑂𝜅²=0,(2.17) where𝑋(𝐻)=−𝜉𝐴(𝐻)=𝜅RPee_𝑛𝑛1−3𝜉𝐻1,sin𝜃Fr^1/𝑛×(𝐻+𝜏)^1/𝑛𝐻−𝜉𝑛×1𝑛+16(𝐻+𝜏)^{(𝑛+1)/𝑛}−𝑛²(2𝑛+1)(3𝑛+1)(𝐻+𝜏)^{(3𝑛+1)/𝑛}+𝜏^{(3𝑛+1)/𝑛}+2𝑛₃𝐻²(2𝑛+1)(3𝑛+1)(4𝑛+1)(𝐻+𝜏)^{(4𝑛+1)/𝑛}−𝜏^{(4𝑛+1)/𝑛}+Re_𝑛𝐻³𝜉(𝑛+1)Pe_𝑛Re^𝑛sin𝜃Fr^{(2−𝑛)/𝑛}×−2𝑛(𝐻+𝜏)^{(2−𝑛)/𝑛}(𝐻−𝑛𝜏)+(𝐻+𝜏)^{(1−𝑛)/𝑛}×(𝐻+𝜏)_{(𝑛+1)/𝑛}(𝑛+1)𝐻−𝑛(𝑛(𝐻++1)𝜏𝐻)^1/𝑛2(𝐻−𝑛𝜏)+𝑛𝜏^{(2𝑛+1)/𝑛}−𝑛2𝑛+1(𝐻+𝜏)^{(2𝑛+1)/𝑛}+𝑛₂−𝜏^{(2𝑛+1)/𝑛}2(𝑛+1)(𝑛+2)𝐻1²(𝐻+𝜏)^{2(𝑛+1)/𝑛}−𝑛−𝜏^{2(𝑛+1)/𝑛}(𝑛+2)𝐻1(𝐻+𝜏)^{(𝑛+2)/𝑛}+(𝐻+𝜏)^2/𝑛(𝐻−𝑛𝜏)−𝑛₂𝐻𝑛+1𝜏𝐻^{(𝑛+1)/𝑛2}(𝐻+𝜏)^{(𝑛+1)/𝑛}−𝜏^{(𝑛+1)/𝑛}+𝜏^{(𝑛+1)/𝑛}𝐵(𝐻)=𝜅Re_𝑛𝐻𝑛(𝐻+𝜏)^1/𝑛,Re^𝑛sin𝜃Fr^{(3−𝑛)/𝑛}(𝐻+𝜏)^1/𝑛×3𝑛²(𝑛+1)(𝑛+2)(2𝑛+1)(𝐻+𝜏)^{2(𝑛+1)/𝑛}𝐻−𝑛(𝑛+1)(𝑛+2)(𝐻+𝜏)^{(𝑛+2)/𝑛}𝐻²+(𝐻+𝜏)^2/𝑛𝐻³−3𝑛³×(𝐻+𝜏)_{(3𝑛+2)/𝑛}(𝑛+1)(𝑛+2)(2𝑛+1)(3𝑛+2)−𝜏^{(3𝑛+2)/𝑛}+2𝑛²(𝑛+1)(2𝑛+1)𝜏^{(2𝑛+1)/𝑛}×𝑛𝑛+1(𝐻+𝜏)^{(𝑛+1)/𝑛}+2𝑛₂−𝜏^{(𝑛+1)/𝑛}−(𝐻+𝜏)^1/𝑛𝐻(𝑛+1)(2𝑛+1)(𝐻+𝜏)^1/𝑛−2𝑛₃𝐻(𝐻+𝜏)^{(2𝑛+1)/𝑛}−𝜏^{(2𝑛+1)/𝑛}(𝑛+1)²(2𝑛+1)(𝐻+𝜏)^{(𝑛+1)/𝑛}+2𝑛₂(𝐻+𝜏)^{(2𝑛+1)/𝑛}−𝜏^{(2𝑛+1)/𝑛}(𝑛+1)²(𝐻+𝜏)^{2(𝑛+1)/𝑛}𝐻−3𝑛𝑛+1(𝐻+𝜏)^{(𝑛+2)/𝑛}𝐻²−𝜅Re_𝑛(𝑛+1)coFsr𝜃Re^𝑛sin𝜃Fr^{(1−𝑛)/𝑛}×(𝐻+𝜏)^1/𝑛(𝐻−𝑛𝜏)𝐻−𝑛𝑛+1(𝐻+𝜏)^{(𝑛+1)/𝑛}+𝑛₂(𝐻−𝑛𝜏)+𝑛𝜏^{(2𝑛+1)/𝑛}(𝑛+1)(2𝑛+1)(𝐻+𝜏)^{(2𝑛+1)/𝑛}𝐶(𝐻)=𝜅₃−𝜏^{(2𝑛+1)/𝑛},Re_𝑛(𝑛+1)Re^𝑛sin𝜃Fr^{(1−𝑛)/𝑛}×Fi_𝑛(𝐻+𝜏)^1/𝑛(𝐻−𝑛𝜏)𝐻−𝑛𝑛+1(𝐻+𝜏)^{(𝑛+1)/𝑛}+𝑛₂(𝐻−𝑛𝜏)+𝑛𝜏^{(2𝑛+1)/𝑛}(𝑛+1)(2𝑛+1)(𝐻+𝜏)^{(2𝑛+1)/𝑛}𝐷(𝐻)=𝜅Re_𝑛−𝜏^{(2𝑛+1)/𝑛},Re^𝑛sin𝜃Fr^{(3−𝑛)/𝑛}×𝑛6𝑛²+19𝑛+11(𝑛+1)²(𝑛+2)(2𝑛+1)(𝐻+𝜏)^{(𝑛+3)/𝑛}𝐻−(7𝑛+15)(𝑛+1)(𝑛+2)(𝐻+𝜏)^3/𝑛𝐻²+3𝑛(𝐻+𝜏)^{(3−𝑛)/𝑛}𝐻³−4𝑛(𝑛+1)(2𝑛+1)𝜏^{(2𝑛+1)/𝑛}+4𝑛(𝐻+𝜏)^{(2−𝑛)/𝑛}𝐻(𝑛+1)(2𝑛+1)(𝐻+𝜏)^{(2−𝑛)/𝑛}−3𝑛₂𝐻(𝐻+𝜏)^{(2𝑛+1)/𝑛}−𝜏^{(2𝑛+1)/𝑛}(𝐻+𝜏)^{(1−𝑛)/𝑛}(𝑛+1)(𝑛+2)(2𝑛+1)(3𝑛+2)(𝐻+𝜏)^{(3𝑛+2)/𝑛}+2𝑛₂−𝜏^{(3𝑛+2)/𝑛}(𝑛+1)²(2𝑛+1)𝜏^{(2𝑛+1)/𝑛}−2𝑛₂(𝐻+𝜏)^{(1−𝑛)/𝑛}(𝐻+𝜏)^{(𝑛+1)/𝑛}−𝜏^{(𝑛+1)/𝑛}(𝑛+1)²(2𝑛+1)(𝐻+𝜏)^2/𝑛−𝜅Re_𝑛(𝐻+𝜏)^{(2𝑛+1)/𝑛}−𝜏^{(2𝑛+1)/𝑛}(𝑛+1)coFsr𝜃Re^𝑛sin𝜃Fr^{(1−𝑛)/𝑛}1𝑛(𝐻+𝜏)^{(1−𝑛)/𝑛}−𝜉𝜅(𝐻−𝑛𝜏)𝐻+(𝐻+𝜏)^1/𝑛𝐻𝐸(𝐻)=𝜅P₃e_𝑛1−3𝜉𝐻1,Re_𝑛(𝑛+1)Re^𝑛sin𝜃Fr^{(1−𝑛)/𝑛}Fi_𝑛1𝑛(𝐻+𝜏)^{(1−𝑛)/𝑛}(𝐻−𝑛𝜏)𝐻+(𝐻+𝜏)^1/𝑛𝐻.(2.18)

3. Stability Analysis

As the variation of the film thickness of the base flow is found to be very small for |𝜅𝐻_𝑥|≪1 using an analysis based on Nusselt assumption, the dimensionless film thickness is expressed as𝐻=1+𝜂(𝑥,𝑡),(3.1) where 𝜂(𝑥,𝑡) is the perturbation of the stationary film thickness. The approximation |𝜅𝐻_𝑥|≪1 gives qualitative results for the constant film thickness assumption at the zeroth order. It is important to note that this constant film thickness approximation with long-wave perturbations are reasonable approximations only for certain segments of weakly condensing and evaporating flows. Substituting for 𝐻(𝑥,𝑡) in (2.18) and retaining terms up to the order of 𝜂³, the evolution equation for 𝜂 is obtained as𝜂_𝑡+𝑋^′𝜂+𝐴𝜂_𝑥+𝐵𝜂_𝑥𝑥+𝐶𝜂_{𝑥𝑥𝑥𝑥}=−𝑋^𝜂²2+𝑋^𝜂³6+𝐴^𝜂+𝐴^𝜂²2𝜂^𝑥+𝐵^𝜂+𝐵^𝜂²2𝜂^𝑥𝑥+𝐶^𝜂+𝐶^𝜂²2𝜂^{𝑥𝑥𝑥𝑥}+𝐷+𝐷^𝜂𝜂^2𝑥+𝐸+𝐸^𝜂𝜂_𝑥𝜂_𝑥𝑥𝑥+𝑂𝜂⁴,(3.2) where the values of 𝑋,𝐴,𝐵,𝐶,𝐷,𝐸, and their derivatives are evaluated at the dimensionless height of the film 𝐻=1. It is worth mentioning here that the analysis based on locally valid unsteady equation has been the subject of several investigations of laminar film condensation [7, 8, 15, 17, 18, 34, 39–42]. Equation (3.2) describes the behavior of finite-amplitude disturbances on the power-law film, and it is used to predict the time-wise behavior of an initially sinusoidal disturbance on the power-law film. It is important to remark that the simplified unsteady equation (3.2) is only locally valid.

In order to understand the flow characteristics and the associated time-dependent properties of power-law liquid film down an inclined plane, typical values of the physical parameters considered by Lin and Hwang [24], Usha and Uma [34], and Pascal and D’Alessio [36] have been used for numerical evaluation and are given byℎ₀𝜌=998Kg/m³,𝜇_𝑛=1.002×10⁻³Pas^𝑛,=10⁻⁴m,𝜎=0.0727N/m,𝛾=0.000611,𝜃=60^∘.(3.3) The temperature at the interface is taken as 𝑇_𝑠=373K and the temperature difference between the wall and the interface as 𝑇_𝑠−𝑇_𝑤=±47K. Under such temperature conditions, the phase change parameter 𝜉 takes values |𝜉|=0.0872 (with phase change) and |𝜉|=0 (without phase change). Shear stress parameter 𝜏 varies between 0.0 and 2.0.

3.1. Linear Stability Analysis

For the linear stability analysis, the nonlinear terms of (3.2) are neglected and the linearized equation𝜂_𝑡+𝑋^′𝜂+𝐴𝜂_𝑥+𝐵𝜂_𝑥𝑥+𝐶𝜂_{𝑥𝑥𝑥𝑥}=0(3.4) is obtained. Assuming the normal mode solution as𝜂=𝑎𝑒^{𝑖(𝑥−𝑑𝑡)}+𝑎𝑒^{−𝑖(𝑥−𝑑𝑡)},(3.5) the complex wave celerity corresponding to linear stability problem is given by𝑑=𝑑_𝑟+𝑖𝑑_𝑖=𝐴+𝑖𝐵−𝐶−𝑋^,(3.6) where 𝑑_𝑟 is the linear wave speed and 𝑑_𝑖 is the linear growth rate of the amplitudes. The flow is in a linearly unstable supercritical condition for 𝑑_𝑖>0 and in a linearly stable subcritical condition for 𝑑_𝑖<0. For 𝑑_𝑖=0, the flow is neutrally stable. Zero linear growth rate of the perturbation (𝑑_𝑖=0) gives raise to a cut-off wave number 𝜅=𝜅_𝑐(𝑛,Re_𝑛,We_𝑛,𝜃,𝜉,Pr_𝑛,𝜏).

Figure 2 shows the neutral stability curve for isothermal, condensate, and evaporating power-law films with the effect of wind stress at the interface. It is observed from Figure 2 that the neutral stability curve plotted from the linear stability analysis separates 𝜅−𝑛 plane into two regions depending on the value of phase change parameter 𝜉 and shear stress parameter 𝜏. Region of linear instability (𝑑_𝑖>0) increases with the increase in the shear stress 𝜏 at the interface and decreases with the increase in the phase change parameter 𝜉. The neutral stability curve for condensate film shows that there exists a critical value for 𝑛, above which condensate power-law fluid film is always stable, and this critical value increases with the increase in shear stress 𝜏 at the interface (Figure 2(d)). It is also observed that condensate power-law fluid film is more stable than the corresponding isothermal and evaporating power-law fluid film in the absence of wind stress.

(a) 𝜏=0

(b) 𝜏=0.2

(c) 𝜏=0.4

(d) 𝜉=0.0872

Figure 2 

Neutral stability curves for isothermal (𝜉=0), condensate (𝜉=0.0872), and evaporating (𝜉=−0.0872) films with the effect of wind stress.

Figure 3 shows the temporal growth rate of power-law fluid given by (3.6). Temporal growth rate of the power-law liquid film increases with the increase in shear stress parameter 𝜏 and decrease in power-law index 𝑛. It is observed that temporal growth rate is lower for a dilatant (𝑛=1.05) condensate (𝜉>0) film when 𝜏=0 and higher for a pseudoplastic (𝑛=0.95) evaporating (𝜉<0) film when 𝜏>0. The results of the linear stability analysis are in good agreement with those of Lin and Hwang [24] (isothermal power-law liquid film for zero shear stress at the interface) and Usha and Uma [34] (power-law condensate/evaporating film for zero shear stress at the interface).

(a) 𝜉=0,𝜏=0

(b) 𝜉=0.0872,𝜏=0

(c) 𝜉=−0.0872,𝜏=0

(d) 𝜉=0,𝜏=0.2

(e) 𝜉=0.0872,𝜏=0.2

(f) 𝜉=−0.0872,𝜏=0.2

(g) 𝜉=0,𝜏=0.4

(h) 𝜉=0.0872,𝜏=0.4

(i) 𝜉=−0.0872,𝜏=0.4

Figure 3 

Growth rate for different values of phase change parameter and wind stress parameter.

3.2. Weakly Nonlinear Stability Analysis

As the perturbed wave grows to a finite amplitude, linear stability theory cannot be used to predict the flow behavior accurately. Therefore, in order to examine whether the finite-amplitude disturbance in the linearly stable region causes instability (subcritical instability) and to investigate whether the subsequent nonlinear evolution of disturbances in the linearly unstable region develops into a new equilibrium state with a finite-amplitude (supercritical stability) or grows to be unstable, the nonlinear stability analysis is employed. The nonlinear stability analysis of (3.2) by the method of multiple scales [43] yields𝐿₀+𝜖𝐿₁+𝜖²𝐿₂𝜖𝜂₁+𝜖²𝜂₂+𝜖³𝜂₃=−𝜖²𝑁₂−𝜖³𝑁₃+𝑜𝜖⁴,(3.7) where𝜂𝜕𝜖,𝑥,𝑥₁,𝑡,𝑡₁,𝑡₂=𝜖𝜂₁+𝜖²𝜂₂+𝜖³𝜂₃,𝑡₁=𝜖𝑡,𝑡₂=𝜖²𝑡,𝑥₁=𝜖𝑥,𝜕𝑡⟶𝜕𝜕𝑡+𝜖𝜕𝜕𝑡₁+𝜖²𝜕𝜕𝑡₂;𝜕𝜕𝑥⟶𝜕𝜕𝑥+𝜖𝜕𝜕𝑥₁𝐿₀,=𝜕𝜕𝑡+𝑋^′+𝐴𝜕𝜕𝑥+𝐵𝜕𝜕𝑥²²+𝐶𝜕⁴𝜕𝑥⁴𝐿₁,=𝜕𝜕𝑡₁+𝐴𝜕𝜕𝑥₁+2𝐵𝜕²𝜕𝑥𝜕𝑥₁+4𝐶𝜕⁴𝜕𝑥³𝜕𝑥₁𝐿₂,=𝜕𝜕𝑡₂+𝐵𝜕²𝜕𝑥₁₂+6𝐶𝜕⁴𝜕𝑥²𝜕𝑥₁₂𝑁₂,=𝑋^2𝜂²¹𝑁₃+𝐴^𝜂₁𝜂_1𝑥+𝐵^𝜂₁𝜂_1𝑥𝑥+𝐶^𝜂₁𝜂_{1𝑥𝑥𝑥𝑥}+𝐷𝜂^21𝑥+𝐸𝜂_1𝑥𝜂_{1𝑥𝑥𝑥},=𝑋^𝜂₁𝜂₂+𝑋^6𝜂³¹+𝐵_+𝐴^𝜂₁𝜂_2𝑥+𝜂_1𝑥𝜂₂+𝜂₁𝜂_1𝑥1+𝐶_𝜂₁𝜂_2𝑥𝑥+2𝜂₁𝜂_1𝑥𝑥1+𝜂_1𝑥𝑥𝜂₂+𝐸𝜂𝜂_11𝑥𝜂_{2𝑥𝑥𝑥𝑥}+4𝜂₁𝜂_{1𝑥𝑥𝑥𝑥1}+𝜂_{1𝑥𝑥𝑥𝑥}𝜂₂+𝐷2𝜂_1𝑥𝜂_2𝑥+2𝜂_1𝑥𝜂_1𝑥1+𝐴_𝜂_{2𝑥𝑥𝑥}+3𝜂_1𝑥𝜂_{1𝑥𝑥𝑥1}+𝜂_{1𝑥𝑥𝑥}𝜂_2𝑥+𝜂_{1𝑥𝑥𝑥}𝜂_1𝑥12𝜂²¹𝜂_1𝑥+𝐵^2𝜂²¹𝜂_1𝑥𝑥+𝐶^2𝜂²¹𝜂_{1𝑥𝑥𝑥𝑥}+𝐷^𝜂₁𝜂^21𝑥+𝐸^𝜂₁𝜂_1𝑥𝜂_{1𝑥𝑥𝑥}.(3.8) The solution of (3.7) at the order 𝑂(𝜖) is obtained by solving 𝐿₀𝜂₁=0 and is in the form𝜂₁=𝑎𝑒^{𝑖(𝑥−𝑑𝑟𝑡)}+𝑎𝑒^{−𝑖(𝑥−𝑑𝑟𝑡)},(3.9) where 𝑎(𝑥₁,𝑡₁,𝑡₂) is the nonlinear amplitude function and 𝑎(𝑥₁,𝑡₁,𝑡₂) is its complex conjugate. The solution of the equation 𝐿₀𝜂₂+𝐿₁𝜂₁=−𝑁₂ at the 𝑂(𝜖²) is in the form𝜂₂=𝑒𝑎²𝑒^{2𝑖(𝑥−𝑑𝑟𝑡)}+𝑒𝑎²𝑒^{−2𝑖(𝑥−𝑑𝑟𝑡)}.(3.10) Using the solutions for 𝜂₁ and 𝜂₂ in the 𝑂(𝜖³) equation given by 𝐿₀𝜂₃+𝐿₁𝜂₂+𝐿₂𝜂₁=−𝑁₃, the equation for the perturbation amplitude 𝑎(𝑥₁,𝑡₁,𝑡₂) is obtained as𝜕𝑎𝜕𝑡₂+𝐷₁𝜕²𝑎𝜕𝑥²¹−𝜖⁻²𝑑_𝑖𝑎+𝐸₁+𝑖𝐹₁𝑎²𝑎=0(3.11) from the secular condition for 𝑂(𝜖³), where𝐷₁=𝐵−6𝐶,𝐹₁=𝑋^−5𝐵^+17𝐶^+4𝐷−10𝐸𝑒_𝑖+𝐴^𝑒𝑟+𝐴^𝐸₁2,=𝑋^−5𝐵^+17𝐶^+4𝐷−10𝐸𝑒_𝑟−𝐴^𝑒𝑖+𝑋^2−23𝐵^+32𝐶^+𝐷^−𝐸^𝑒_𝑟,+𝑖𝑒_𝑖=−𝑋^/2+𝐵^−𝐶^+𝐷−𝐸−𝑖𝐴^16𝐶−4𝐵+𝑋^.(3.12) The weakly nonlinear behavior of the fluid film can be investigated using (3.12). It is important to note that such an expansion is only valid for wave numbers close to neutral and not near critical when 𝜅 approaches zero. The solution of (3.12) for a filtered wave in which spatial modulation does not exist and the diffusion terms in (3.12) vanishes is obtained by taking 𝑎=𝑎₀𝑒^{−𝑖𝑏(𝑡2)𝑡2}. This leads to the Ginzburg-Landau equation given by𝜕𝑎₀𝜕𝑡₂𝜕=𝑏𝜖𝑡⁻²𝑑_𝑖−𝐸₁𝑎²⁰𝑎₀,(3.13)₂𝑡₂𝜕𝑡₂=𝐹₁𝑎²⁰.(3.14) The second term in (3.13) induced by the effect of nonlinearity can either accelerate or decelerate the exponential growth of the linear disturbance depending upon the signs of 𝑑_𝑖 and 𝐸₁. The perturbed wave speed caused by the infinitesimal disturbances appearing in the nonlinear system can be modified using (3.13). The threshold amplitude 𝜖𝑎₀ is given by𝜖𝑎₀=𝑑_𝑖𝐸₁,(3.15) and the nonlinear wave speed is given as𝑁𝑐_𝑟=𝜖²𝑏=𝑑_𝑟+𝑑_𝑖𝐹₁𝐸₁.(3.16) It is observed from (3.15) that in the linearly unstable region (𝑑_𝑖>0), the condition for existence of a supercritical stable region is 𝐸₁>0 and 𝜖𝑎₀ is the threshold amplitude. In the linearly stable region (𝑑_𝑖<0), if 𝐸₁<0, then the flow has the behavior of subcritical instability and 𝜖𝑎₀ is the threshold amplitude. The condition for the existence of a subcritical stable region is 𝐸₁>0, and 𝐸₁=0 gives the condition of existence of a neutral stability curve.

The neutral stability curves are obtained from (3.6) and (3.12) by equating to zero, the linear amplification rate 𝑑_𝑖, and the nonlinear amplification rate 𝐸₁. Figure 4 shows the regions of subcritical stability (𝑑_𝑖<0,𝐸₁>0) and subcritical instability (𝑑_𝑖<0,𝐸₁<0) in the linearly stable region and supercritical stability (𝑑_𝑖>0,𝐸₁>0) and supercritical explosive state (𝑑_𝑖>0,𝐸₁<0) in the linearly unstable region. It is observed that subcritical stable region (𝑑_𝑖<0,𝐸₁>0) decreases, and supercritical stability (𝑑_𝑖>0,𝐸₁>0) increases with the increase in shear stress parameter 𝜏. However, for the condensate film (𝜉>0), region of supercritical explosive state (𝑑_𝑖>0,𝐸₁<0) and supercritical stability (𝑑_𝑖>0,𝐸₁>0) exists only for the pseudoplastic fluids (𝑛<1) which increases with the increase in 𝜏.

(a) 𝜉=0,𝜏=0

(b) 𝜉=0.0872,𝜏=0

(c) 𝜉=−0.0872,𝜏=0

(d) 𝜉=0,𝜏=0.2

(e) 𝜉=0.0872,𝜏=0.2

(f) 𝜉=−0.0872,𝜏=0.2

(g) 𝜉=0,𝜏=0.4

(h) 𝜉=0.0872,𝜏=0.4

(i) 𝜉=−0.0872,𝜏=0.4

Figure 4 

Neutral stability curves in the 𝜅-𝑛 plane for different values of phase change parameter 𝜉 (𝜉=0—isothermal film flow; 𝜉=0.0872—condensate film flow; 𝜉=−0.0872—evaporating film flow) and wind stress parameters 𝜏.

Figures 5 and 6 show the threshold amplitude and nonlinear wave speed in the supercritical stable region for various values of phase change parameter 𝜉 and shear stress parameter 𝜏. The flow system is stable if the finite amplitude of the disturbance is more than the threshold amplitude and explosively unstable otherwise. The results show that the threshold amplitude and nonlinear wave speed increases with increase in 𝜏 and decreases with increase in 𝜉 and 𝑛.

(a) 𝜉=0,𝜏=0

(b) 𝜉=0.0872,𝜏=0

(c) 𝜉=−0.0872,𝜏=0

(d) 𝜉=0,𝜏=0.2

(e) 𝜉=0.0872,𝜏=0.2

(f) 𝜉=−0.0872,𝜏=0.2

(g) 𝜉=0,𝜏=0.4

(h) 𝜉=0.0872,𝜏=0.4

(i) 𝜉=−0.0872,𝜏=0.4

Figure 5 

Threshold amplitude in the supercritical stable region.

(a) 𝜉=0,𝜏=0

(b) 𝜉=0.0872,𝜏=0

(c) 𝜉=−0.0872,𝜏=0

(d) 𝜉=0,𝜏=0.2

(e) 𝜉=0.0872,𝜏=0.2

(f) 𝜉=−0.0872,𝜏=0.2

(g) 𝜉=0,𝜏=0.4

(h) 𝜉=0.0872,𝜏=0.4

(i) 𝜉=−0.0872,𝜏=0.4

Figure 6 

Nonlinear wave speed in the supercritical stable region.

The above weakly nonlinear analysis shows the existence of both a supercritical stable region and a subcritical unstable region for isothermal and nonisothermal power-law liquid film down an inclined plane in the presence of shear stress at the interface. It is observed that the evolution of two-dimensional waves depends strongly on the initial disturbance wave number. There exists a value of 𝜅_𝑠 (obtained from equating denominator of 𝐸₁=0) such that, when 𝜅_𝑠<𝜅<𝜅_𝑐 (obtained from 𝑑_𝑖=0; cut-off wave number), the flow is supercritically stable and nonlinear equilibration occurs after the initial instability. The weakly nonlinear stability analysis shows that 𝜅_𝑠=𝜅_𝑐/2, and the curve 𝜅=𝜅_𝑠 separates the linearly unstable region into two portions where the nonlinear waves attain a finite equilibrium amplitude (𝑑_𝑖>0,𝐸₁>0) or reach an explosive state (𝑑_𝑖>0,𝐸₁<0), and the flow is supercritically equilibrated.

3.3. Nonlinear Analysis

In order to understand the mechanism responsible for the transfer of energy from the basic state to the disturbance, the evolution of finite-amplitude perturbations is considered. On the evolution of these perturbations, the influence of wind stress and mass transfer at the interface of a thin power-law liquid film down an inclined plane is examined by numerically solving the nonlinear evolution equation. The initial disturbance is taken to be a sinusoidal wave with small amplitude given by𝐻(𝑥,0)=1−0.1cos(𝜅𝑥).(3.17) The evolution of waves with time is obtained by solving the evolution equation (2.18) by using Lee’s three-time level [44] finite-difference method in a periodic domain [−𝜋/𝜅,𝜋/𝜅]. The finite-difference scheme achieves linearity in the unknown 𝐻_𝑖,𝑗+1 by evaluating all coefficients of 𝐻_𝑖,𝑗+1 at a time level of known solution values, preserves stability by averaging 𝐻_𝑖,𝑗 over three time levels, and maintains accuracy by using central difference approximations, where 𝑗 denotes the number of time steps and 𝑖 is a spatial grid point. Special attention is given to the discretization of the wave propagation term so as to overcome the solution blowup for very small initial wave numbers. The scheme possesses both conservative and transportive properties and allows to follow the solution over larger periods of time. The resulting algebraic equations are solved using MATLAB, and the iterations are continued until the local minimum layer thickness becomes smaller than the maximum error bound of 10⁻¹⁰. The computations are performed with Δ𝑡=0.01 and with the number of nodes along the spatial direction as 𝑁=100 and Δ𝑥=2𝜋/𝑁𝜅. It is to be noted that the solution obtained with a decrease in Δ𝑡 or an increase in 𝑁 does not show any deviation from the values obtained for Δ𝑡=0.01, Δ𝑥=2𝜋/𝑁𝜅, 𝑁=100. Further, it is confirmed that the wave that emerges on the film surface does not change its shape for a relatively long period of time.

The weakly nonlinear analysis predicts the occurrence of a wave number 𝜅_𝑠 that separates the regimes of supercritical 𝜅_𝑠<𝜅<𝜅_𝑐 and subcritical 𝜅<𝜅_𝑠 domains (Figure 4). For a wave number 𝜅 close to 𝜅_𝑐, the solution may evolve into a stable, almost sinusoidal wave of small finite amplitude. For 𝜅 close to 𝜅_𝑚=𝜅_𝑐/√2, the surface wave approaches the form of a solitary wave as observed in experimental results by Alekseenko et al. [45] and Liu et al. [46]. The evolution of the film in the supercritical stable region is analyzed by performing numerical simulations for initial disturbance wave number 𝜅=𝜅_𝑐,𝜅_𝑚(=𝜅_𝑐/√2), and 𝜅_𝑠(=𝜅_𝑐/2). The evolution of the film is examined for values of 𝜉 in the range −0.0872 to 0.0872 and 𝜏 between 0 and 2.0.

Figures 7–9 show the evolution of the free surface configuration of isothermal (𝜉=0) pseudoplastic (𝑛=0.95), Newtonian (𝑛=1.0), and dilatant (𝑛=1.05) fluids at various instants of time for different values of shear stress parameter at the interface. Wave number is chosen in the range 𝜅_𝑠(=𝜅_𝑐/2)<𝜅<𝜅_𝑐 (Figure 7—𝜅=0.06, 𝜅_𝑐=0.0868 when 𝜏=0,𝜅_𝑐=0.0971 when 𝜏=0.2, and 𝜅_𝑐=0.1064 when 𝜏=0.4; Figure 8—𝜅=0.05,𝜅_𝑐=0.0569 when 𝜏=0,𝜅_𝑐=0.0635 when 𝜏=0.2, and 𝜅_𝑐=0.0694 when 𝜏=0.4; Figure 9—𝜅=0.03, 𝜅_𝑐=0.0365 when 𝜏=0,𝜅_𝑐=0.0409 when 𝜏=0.2, and 𝜅_𝑐=0.0449 when 𝜏=0.4). It is observed that the growth rate is much more important and the distortion of the free surface is conspicuous with increase in time. The wave amplitude decays after reaching a maximum, and the rear becomes longer and longer to attain a one-hump solitary-like wave at different instant of time for various values of 𝜏. The large amplitude wave is then followed by a small amplitude capillary wave. In all the cases, it is observed that finite-amplitude permanent wave emerges. For a given 𝑛, amplitude of the wave increases with the increase in shear stress parameter 𝜏. The corresponding evolution of maximum (𝐻_max) and minimum (𝐻_min) thickness is presented in Figure 10. The disturbance amplitude 𝐻_max remains constant for initial times 𝑡∼0-5 for pseudoplastic fluids (𝑛=0.95), 𝑡∼0-10 for Newtonian fluids (𝑛=1.0), and 𝑡∼0-20 for dilatant fluids (𝑛=1.05) and then increases monotonically to attain a maximum amplitude. Isothermal pseudoplastic fluid attains the maximum amplitude in short time compared to Newtonian and dilatant fluid, and this amplitude further increases with the presence of shear stress at the interface. The distortion of the free surface is more significant when the effect of shear stress increases (Figure 11).

(a)

(b)

(c)

Figure 7 

Evolution of free surface for an isothermal (𝜉=0) pseudoplastic (𝑛=0.95) film when 𝜅=0.06 and 𝜏=0,0.2,0.4.

(a)

(b)

(c)

Figure 8 

Evolution of free surface for an isothermal (𝜉=0) Newtonian (𝑛=1.0) film when 𝜅=0.05 and 𝜏=0,0.2,0.4.

(a)

(b)

(c)

Figure 9 

Evolution of free surface for isothermal (𝜉=0) dilatant (𝑛=1.05) film when 𝜅=0.03 and 𝜏=0,0.2,0.4.

(a) 𝑛=0.95,𝜉=0.0

(b) 𝑛=1.0,𝜉=0.0

(c) 𝑛=1.05,𝜉=0.0

Figure 10 

Maximum and minimum amplitude of the waves down an inclined plane for an isothermal power-law film.

(a)

(b)

(c)

Figure 11 

Evolution of free surface for an isothermal film (𝜉=0) when 𝜏=2.0.

Weakly nonlinear stability analysis predicted that for the condensate film (𝜉>0), region of supercritical stability (𝑑_𝑖>0,𝐸₁>0) exists only for the pseudoplastic fluids (𝑛=0.95). This is the case when the saturated vapor temperature is more than the wall temperature. When condensation is considered, the static layer grows thicker with time. Evolution of condensate pseudoplastic liquid film down an inclined plane is depicted in Figure 12 for 𝜉=0.01,0.02,0.04,and0.06 and 𝜏=0. The result shows that pseudoplastic film down an inclined plane thickens faster as the phase change parameter 𝜉 increases gradually in the absence of shear stress parameter 𝜏. The effect of wind stress on the interface of the pseudoplastic film takes little longer time to thicken (Figure 13) and makes the film more corrugated as 𝜏 increases (Figures 13(a)(ii), 13(b)(ii), 13(c)(ii)).

(a) 𝑛=0.95,𝜉=0.01

(b) 𝑛=0.95,𝜉=0.02

(c) 𝑛=0.95,𝜉=0.04

(d) 𝑛=0.95,𝜉=0.06

Figure 12 

Evolution of free surface for condensating (𝜉>0) pseudoplastic (𝑛=0.95) film when 𝜅=0.06 and 𝜏=0.

(a)

(b)

(c)

Figure 13 

Evolution of free surface for condensating (𝜉=0.0872) pseudoplastic (𝑛=0.95) film when 𝜅=0.06 and 𝜏=0,0.2,0.4.

Figure 14 shows the evolution of the free-surface shape for evaporating pseudoplastic, Newtonian, and dilatant fluids when 𝜉=−0.01 and −0.0872 with a zero shear stress at the interface. This is the case when the saturated vapor temperature is less than the wall temperature. It is observed that as 𝑛 increases, depth of the liquid layer disappears in a finite time 𝑡=𝑡_𝑟 decreases leading to rupture instability. The rupture instability is thus augmented by the evaporative effect. Plot describes that the dilatant film evaporates faster than the pseudoplastic and Newtonian film. For a given 𝑛 (given pseudoplastic, Newtonian, or dilatant fluid), evaporation of the film happens at earlier instant of time when the phase change parameter 𝜉 decreases (Figure 14). Presence of wind stress on the interface does not influence the time of rupture 𝑡_𝑟 for any 𝑛 whereas it has little influence on the position where the rupture takes place (Figure 15).

(a) 𝑛=0.95,𝜉=−0.01

(b) 𝑛=1.0,𝜉=−0.01

(c) 𝑛=1.05,𝜉=−0.01

(d) 𝑛=0.95,𝜉=−0.0872

(e) 𝑛=1.0,𝜉=−0.0872

(f) 𝑛=1.05,𝜉=−0.0872

Figure 14 

Evolution of free surface for an evaporating film (𝜉<0) when 𝜅=0.06 and 𝜏=0.

(a) 𝑛=0.95,𝜏=0.2

(b) 𝑛=1.0,𝜏=0.2

(c) 𝑛=1.05,𝜏=0.2

(d) 𝑛=0.95,𝜏=2.0

(e) 𝑛=1.0,𝜏=2.0

(f) 𝑛=1.05,𝜏=2.0

Figure 15 

Evolution of free surface for an evaporating film (𝜉=−0.00872) when 𝜅=0.06 and 𝜏=0.2,2.0.

4. Conclusion

The influence of prescribed cocurrent superficial shear stress on the dynamics and stability of a condensate or evaporating power-law liquid film falling down an inclined plane has been analyzed by the method of long-wave perturbation. The interfacial boundary conditions include the effects of phase change across the interface. The evolution equation of the Benney type incorporating the effect of superficial shear stress on the surface of a nonisothermal power-law film has been derived. As the power-law exponent “𝑛” decreases, the effective viscosity decreases, and, hence, “𝑛” influences the Reynolds number Re_𝑛, Weber number We_𝑛, and Prandtl number Pr_𝑛. The results of the linear stability analysis reveal that the effect of decreasing the phase change parameter 𝜉 or increasing the shear stress parameter 𝜏 is to destabilize the film flow system. Further, the dimensional quantities used to discuss the stability characteristics of the power-law model in terms of the power-law exponent “𝑛” show that there exists a critical value of “𝑛” for the condensate film, above which the film flow system is always stable (Figure 2), and this critical value of “𝑛” increases with increase in shear stress parameter 𝜏 (Figure 2(d)), while isothermal and evaporating power-law fluid films are unstable for any value of power-law index “𝑛”. The weakly nonlinear stability analysis of the power-law film flow system using long-wave theory is even more qualitative than the linearized stability results, and it reveals that both subcritical instability (𝑑_𝑖<0,𝐸₁<0) and supercritical stability (𝑑_𝑖>0,𝐸₁>0) are possible for isothermal and evaporating power-law fluid films and supercritical stability (𝑑_𝑖>0,𝐸₁>0) and supercritical explosive state (𝑑_𝑖>0,𝐸₁<0) are not possible for condensate dilatant fluid films (Figure 4).

The nonlinear evolution of film thickness is found by numerically solving the evolution equation in a periodic domain. The numerical simulations confirm the results found on the basis of linear and weakly nonlinear stability analysis. For an isothermal power-law liquid film, the finite amplitude waves are stable solutions of the evolution equation when the wave numbers are less than the cut-off wave number 𝜅_𝑐.The permanent waves are nearly sinusoidal for initial wave numbers close to the cut-off wave number, and, for values much smaller than the cut-off value, the permanent waves are of solitary type. It is interesting to note that the range of supercritical stability broadens with an increase in the shear stress parameter 𝜏 for fixed values of other nondimensional parameters governing the film flow system. The nonlinear interactions in falling isothermal power-law films down an inclined plane exhibit a tendency toward permanent two-dimensional waves for very thin films at small wave numbers close to the cut-off wave number. The nonzero shear stress at the interface promotes the growth rate of the wave amplitude (Figure 10).

The results reveal that (i) for an isothermal dilatant (pseudoplastic) liquids, the maximum wave amplitude is always smaller (larger) than that for a Newtonian liquid and the amplitude of permanent wave increases with the increase in prescribed shear stress 𝜏 at the interface in cocurrent direction; (ii) condensation of pseudoplastic film happens for the earlier instant of time when the phase change parameter increases and the effect of shear stress at the interface makes the film more corrugated; (iii) dilatant (pseudoplastic) evaporating liquid film attains rupture faster (slower) than that of Newtonian liquid film, and the prescribed shear stress 𝜏 at the interface does not influence the time of rupture and has little influence on the position where the rupture takes place.

Appendix

The zeroth- and the first-order solutions of the physical quantities are given by 𝑢₀=Re^𝑛sin𝜃Fr^1/𝑛𝑛+𝑛1(𝐻+𝜏)^{(𝑛+1)/𝑛}𝑣₀−(𝐻−𝑦+𝜏)^{(𝑛+1)/𝑛},=−Re^𝑛sin𝜃Fr^1/𝑛𝐻_𝑥(𝐻+𝜏)^1/𝑛𝑦+𝑛𝑛+1(𝐻−𝑦+𝜏)^{(𝑛+1)/𝑛}−𝑛𝑝₀𝑛+1(𝐻+𝜏)^{(𝑛+1)/𝑛},=cos𝜃Fr(𝐻−𝑦)−𝜅²Θ₀Fi_𝑛𝐻_𝑥𝑥,=𝑦𝐻,(A.1)𝑢₁=−Re^𝑛𝜉(𝑛+1)𝜅Pe_𝑛Re^𝑛sin𝜃Fr^{(2−𝑛)/𝑛}1×(𝐻+𝜏)_1/𝑛𝐻+𝑛𝜏_{(𝑛+1)/𝑛}(𝐻+𝜏)^1/𝑛(𝐻−𝑛𝜏)−(𝐻−𝑦+𝜏)^1/𝑛(𝐻−𝑦−𝑛𝜏)−𝑛(𝐻+𝜏)^1/𝑛−(𝐻−𝑦+𝜏)^1/𝑛𝑛+2(𝐻+𝜏)^{(𝑛+2)/𝑛}+Re_𝑛−(𝐻−𝑦+𝜏)^{(𝑛+2)/𝑛}Re^𝑛sin𝜃Fr^{(3−𝑛)/𝑛}𝐻_𝑥(𝐻+𝜏)^1/𝑛×−𝑛(𝑛+1)(𝑛+2)(𝐻+𝜏)^{(𝑛+2)/𝑛}+𝑛𝐻−(𝐻−𝑦+𝜏)^{(𝑛+2)/𝑛}(𝐻−𝑦)₂(4𝑛+5)2(𝑛+1)²(𝑛+2)(2𝑛+1)(𝐻+𝜏)^{2(𝑛+1)/𝑛}−2𝑛₂−(𝐻−𝑦+𝜏)^{2(𝑛+1)/𝑛}(𝑛+1)(2𝑛+1)𝜏^{(2𝑛+1)/𝑛}(𝐻+𝜏)^1/𝑛−(𝐻+𝜏)_1/𝑛−(𝐻−𝑦+𝜏)^1/𝑛(𝐻−𝑦+𝜏)^1/𝑛(𝐻−𝑦)𝐻+(𝐻+𝜏)^2/𝑛𝐻²−𝑛𝑛+1(𝐻+𝜏)^{(𝑛+2)/𝑛}−𝑛𝐻−(𝐻+𝜏)^1/𝑛(𝐻−𝑦+𝜏)^{(𝑛+1)/𝑛}𝐻𝑛+1(𝐻+𝜏)^{(𝑛+2)/𝑛}+𝑛𝐻−(𝐻+𝜏)^{(𝑛+1)/𝑛}(𝐻−𝑦+𝜏)^1/𝑛(𝐻−𝑦)𝑛+1²(𝐻+𝜏)^{2(𝑛+1)/𝑛}+Re_𝑛−(𝐻+𝜏)^{(𝑛+1)/𝑛}(𝐻−𝑦+𝜏)^{(𝑛+1)/𝑛}(𝑛+1)Re^𝑛sin𝜃Fr^{(1−𝑛)/𝑛}−coFsr𝜃𝐻^𝑥×(𝐻+𝜏)_1/𝑛+𝜅²Fi_𝑛𝐻_𝑥𝑥𝑥𝑣(𝐻−𝑛𝜏)−(𝐻−𝑦+𝜏)^1/𝑛(𝐻−𝑦−𝑛𝜏)₁,=−Re^𝑛𝜉(𝑛+1)𝜅Pe_𝑛Re^𝑛sin𝜃Fr^{(2−𝑛)/𝑛}𝐻_𝑥𝐻²×𝑛𝑛+22(𝑛𝑛+1)(𝐻+𝜏)^{2(𝑛+1)/𝑛}+(𝐻+𝜏)_1/𝑛−(𝐻−𝑦+𝜏)^{2(𝑛+1)/𝑛}−(𝐻+𝜏)^{(𝑛+2)/𝑛}𝑦−𝑛+𝑛1(𝐻+𝜏)^{(𝑛+1)/𝑛}+(𝐻+𝜏)_1/𝑛(𝐻−𝑛𝜏)−(𝐻−𝑦+𝜏)^{(𝑛+1)/𝑛}(𝐻−𝑦−𝑛𝜏)+𝑛(𝐻−𝑛𝜏)𝑦₂(𝑛+1)(2𝑛+1)(𝐻+𝜏)^{(2𝑛+1)/𝑛}−𝑛𝜏_{(𝑛+1)/𝑛}−(𝐻−𝑦+𝜏)^{(2𝑛+1)/𝑛}−Re_𝑛𝑛+𝑛1(𝐻+𝜏)^{(𝑛+1)/𝑛}−(𝐻−𝑦+𝜏)^{(𝑛+1)/𝑛}−(𝐻+𝜏)^1/𝑛𝑦𝜉(𝑛+1)𝜅Pe_𝑛Re^𝑛sin𝜃Fr^{(2−𝑛)/𝑛}𝐻_𝑥×−𝑛𝐻𝑛+2(𝐻+𝜏)^{(𝑛+2)/𝑛}−(𝐻−𝑦+𝜏)^{(𝑛+2)/𝑛}−2𝑛(𝐻+𝜏)^{(2−𝑛)/𝑛}+(𝐻+𝜏)_1/𝑛(𝐻−𝑛𝜏)𝑦+𝜏_{(𝑛+1)/𝑛}(𝐻+𝜏)^1/𝑛(𝐻−𝑛𝜏)−(𝐻−𝑦+𝜏)^1/𝑛(𝐻−𝑦−𝑛𝜏)+(𝐻+𝜏)𝑛(𝐻_{(1−𝑛)/𝑛}+𝜏)^1/𝑛−(𝐻−𝑦+𝜏)^1/𝑛−(𝐻+𝜏)^{(1−𝑛)/𝑛}𝑦𝑛+1(𝐻+𝜏)^{(𝑛+1)/𝑛}−𝑛(𝐻−𝑛𝜏)−(𝐻−𝑦+𝜏)^{(𝑛+1)/𝑛}(𝐻−𝑦−𝑛𝜏)2𝑛+1(𝐻+𝜏)^{(2𝑛+1)/𝑛}−Re_𝑛−(𝐻−𝑦+𝜏)^{(2𝑛+1)/𝑛}Re^𝑛sin𝜃Fr^{(3−𝑛)/𝑛}𝐻^𝑥𝑥(𝐻+𝜏)^1/𝑛+𝐻^2𝑥×𝑛𝑛(𝐻+𝜏)^{(1−𝑛)/𝑛}(𝑛+1)(𝑛+2)2(𝑛𝑛+1)(𝐻+𝜏)^{2(𝑛+1)/𝑛}−(𝐻+𝜏)_{(𝑛+2)/𝑛}𝐻−(𝐻−𝑦+𝜏)^{2(𝑛+1)/𝑛}(𝐻−𝑦)−𝑛𝐻𝑦₂2(𝑛+1)(3𝑛+2)(𝐻+𝜏)^{(3𝑛+2)/𝑛}−𝑛₂−(𝐻−𝑦+𝜏)^{(3𝑛+2)/𝑛}(4𝑛+5)2(𝑛+1)²(𝑛+2)(2𝑛+1)3𝑛𝑛+2(𝐻+𝜏)^{(3𝑛+2)/𝑛}−(𝐻+𝜏)_{2(𝑛+1)/𝑛}−(𝐻−𝑦+𝜏)^{(3𝑛+2)/𝑛}+2𝑛₂𝑦(𝑛+1)(2𝑛+1)𝜏^{(2𝑛+1)/𝑛}𝑛+𝑛1(𝐻+𝜏)^{(𝑛+1)/𝑛}+(𝐻+𝜏)_1/𝑛−(𝐻−𝑦+𝜏)^{(𝑛+1)/𝑛}−(𝐻+𝜏)^1/𝑛𝑦𝐻−𝑛𝑛+1(𝐻+𝜏)^{(𝑛+1)/𝑛}+2𝑛₂𝐻−(𝐻−𝑦+𝜏)^{(𝑛+1)/𝑛}(𝐻−𝑦)(𝑛+1)(2𝑛+1)(𝐻+𝜏)^{(2𝑛+1)/𝑛}+(𝐻+𝜏)_2/𝑛−(𝐻−𝑦+𝜏)^{(2𝑛+1)/𝑛}𝐻²𝑦−2𝑛𝑛+1(𝐻+𝜏)^{(𝑛+2)/𝑛}𝐻𝑦+𝑛𝑛+1²+𝑛(𝐻+𝜏)^{2(𝑛+1)/𝑛}𝑦𝑛+1(𝐻+𝜏)^{(𝑛+1)/𝑛}𝑛+𝑛1(𝐻+𝜏)^{(𝑛+1)/𝑛}−𝑛𝐻−(𝐻−𝑦+𝜏)^{(𝑛+1)/𝑛}(𝐻−𝑦)₂(𝑛+1)(2𝑛+1)(𝐻+𝜏)^{(2𝑛+1)/𝑛}−𝑛−(𝐻−𝑦+𝜏)^{(2𝑛+1)/𝑛}𝑛+1²𝑛2𝑛+1(𝐻+𝜏)^{(𝑛+1)/𝑛}(𝐻+𝜏)^{(2𝑛+1)/𝑛}−Re_𝑛−(𝐻−𝑦+𝜏)^{(𝑛+1)/𝑛}Re^𝑛sin𝜃Fr^{(3−𝑛)/𝑛}𝐻^2𝑥(𝐻+𝜏)^1/𝑛×2𝑛(𝑛+1)(2𝑛+1)−2(𝑛𝑛+1)(𝐻+𝜏)^{2(𝑛+1)/𝑛}+(𝐻+𝜏)_{(𝑛+2)/𝑛}−(𝐻−𝑦+𝜏)^{2(𝑛+1)/𝑛}+𝑛𝑦𝑛+1𝑛+𝑛2(𝐻+𝜏)^{(𝑛+2)/𝑛}−𝑛₂𝐻−(𝐻−𝑦+𝜏)^{(𝑛+2)/𝑛}(𝐻−𝑦)−(𝐻+𝜏)^2/𝑛𝑦2(𝑛+1)(𝑛+2)(𝐻+𝜏)^{2(𝑛+1)/𝑛}+2𝑛−(𝐻−𝑦+𝜏)^{2(𝑛+1)/𝑛}(𝑛+1)(2𝑛+1)𝜏^{(2𝑛+1)/𝑛}𝑛(𝐻+𝜏)^1/𝑛+1−(𝐻−𝑦+𝜏)^1/𝑛−(𝐻+𝜏)^{(1−𝑛)/𝑛}𝑦𝑛(𝐻+𝜏)^{(1−𝑛)/𝑛}𝐻−𝑛𝑛+1(𝐻+𝜏)^{(𝑛+1)/𝑛}+𝑛𝐻−(𝐻−𝑦+𝜏)^{(𝑛+1)/𝑛}(𝐻−𝑦)₂(𝑛+1)(2𝑛+1)(𝐻+𝜏)^{(2𝑛+1)/𝑛}+1−(𝐻−𝑦+𝜏)^{(2𝑛+1)/𝑛}𝑛(𝐻+𝜏)^1/𝑛𝐻+−𝑛𝑛(𝐻+𝜏)^1/𝑛𝐻−(𝐻−𝑦+𝜏)^1/𝑛(𝐻−𝑦)₂𝑛+1(𝐻+𝜏)^{(𝑛+1)/𝑛}+𝑛−(𝐻−𝑦+𝜏)^{(𝑛+1)/𝑛}(𝑛+1)(2𝑛+1)(𝐻+𝜏)^{(1−𝑛)/𝑛}+1𝐻(𝐻+𝜏)^{(2𝑛+1)/𝑛}−(𝐻−𝑦+𝜏)^{(2𝑛+1)/𝑛}𝑛+1(𝐻+𝜏)^{(𝑛+1)/𝑛}𝑛(𝐻+𝜏)^1/𝑛−𝑛𝐻−(𝐻−𝑦+𝜏)^1/𝑛(𝐻−𝑦)₂𝑛+1(𝐻+𝜏)^{(𝑛+1)/𝑛}+2−(𝐻−𝑦+𝜏)^{(𝑛+1)/𝑛}𝑛(𝐻+𝜏)^{(2−𝑛)/𝑛}𝐻²𝑦−2𝑛+1(𝐻+𝜏)^2/𝑛+Re_𝑛𝐻𝑦(𝑛+1)Re^𝑛sin𝜃Fr^{(1−𝑛)/𝑛}coFsr𝜃𝐻^𝑥𝑥×(𝐻+𝜏)_1/𝑛−𝜅²Fi_𝑛𝐻_{𝑥𝑥𝑥𝑥}−𝑛(𝐻−𝑛𝜏)𝑦𝑛+1(𝐻+𝜏)^{(𝑛+1)/𝑛}+𝑛(𝐻−𝑛𝜏)−(𝐻−𝑦+𝜏)^{(𝑛+1)/𝑛}(𝐻−𝑦−𝑛𝜏)₂(𝑛+1)(2𝑛+1)(𝐻+𝜏)^{(2𝑛+1)/𝑛}+Re_𝑛−(𝐻−𝑦+𝜏)^{(2𝑛+1)/𝑛}(𝑛+1)Re^𝑛sin𝜃Fr^{(1−𝑛)/𝑛}coFsr𝜃𝐻^2𝑥×1−𝜅²Fi_𝑛𝐻_𝑥𝐻_𝑥𝑥𝑥𝑛(𝐻+𝜏)^{(1−𝑛)/𝑛}−(𝐻+𝜏)_1/𝑛(𝐻−𝑛𝜏)𝑦+(𝐻+𝜏)^1/𝑛𝑦Θ(𝐻−𝑛𝜏)−(𝐻−𝑦+𝜏)^1/𝑛(𝐻−𝑦−𝑛𝜏)₁,=Pe_𝑛6𝜅Pe𝜉_𝑛𝐻³𝐻³−𝑦³−6𝜅P𝜉e_𝑛𝐻(𝐻−𝑦)+Re^𝑛sin𝜃Fr^1/𝑛𝑛𝑛+1𝐻1²𝐻_𝑥×(𝐻+𝜏)^{(𝑛+1)/𝑛}(𝐻−6𝑦)³−𝐻²(𝐻−𝑦)−𝑛₂6(2𝑛+1)(3𝑛+1)(𝐻−𝑦+𝜏)^{(3𝑛+1)/𝑛}+2𝑛₃−(𝐻+𝜏)^{(3𝑛+1)/𝑛}(𝐻−𝑦)×(2(𝐻𝑛+−1𝑦)+(3𝑛𝜏)+1)(4_{(4𝑛+1)/𝑛}𝑛+1)+𝑦𝐻−1(𝐻+𝜏)^{(4𝑛+1)/𝑛}−𝑦𝐻𝜏^{(4𝑛+1)/𝑛}.(A.2)