Open access peer-reviewed chapter

On the Thermodynamic Consistency of a Two Micro-Structured Thixotropic Constitutive Model

Written By

Hilbeth P. Azikri de Deus and Mikhail Itskov

Submitted: 22 November 2017 Reviewed: 26 February 2018 Published: 05 November 2018

DOI: 10.5772/intechopen.75987

From the Edited Volume

Selected Problems of Contemporary Thermomechanics

Edited by Jerzy Winczek

Chapter metrics overview

1,038 Chapter Downloads

View Full Metrics

Abstract

The time-dependent rheological behavior of the thixotropic fluids is presented in various industrial fields (cosmetics, food, oil, etc.). Usually, a couple of equations define constitutive model for thixotropic substances: a constitutive equation based on linear viscoelastic models and a rate equation (an equation related to the micro-structural evolution of the substance). Many constitutive models do not take into account the micro-structural dependence of the shear modulus and viscosity in the dynamic principles from which are developed. The modified Jeffreys model (considering only one single micro-structure type) does not show this incoherence in its formulation. In this chapter, a constitutive model for thixotropic fluids, based on modified Jeffreys model, is presented with the addition of one more micro-structure type, besides of comments on some possible generalizations. The rheological coherence of this constitutive model and thermodynamic consistency are analyzed too. This model takes into account a simple isothermal laminar shear flows, and the micro-structures dynamics are relate to Brownian motion and de Gennes Reptation model via the Smoluchowski™s coagulation theory.

Keywords

  • thixotropic fluids
  • thermodynamic consistency
  • modified Jeffreys model

1. Introduction

The thixotropic substances are considered structured fluids and have a rheological behavior (time-dependent) guided by their structural nature [15]. In terms of models (constitutive systems), this behavior is represented by a couple of time-dependent equations. These equations connect the micro-structural characteristics to the rheological behavior. In many works, this system of equations is based on a qualitative way, without any formal justification on the physical principles. In this chapter, a different approach is presented based on some well-established physical principles (Smoluchowski’s coagulation theory [610], Brownian motion [69] and de Gennes reptation model [8, 11, 12]).

The rheological properties evolution, in many thixotropy models [2, 5, 1317], is presented, generically, in terms of a couple of equations, as follows:

τ=τλγ̇,constitutive relationship basedonlinear viscoelastic models;E1
λ̇=Gγ̇λγ̇,relation associated to the microstructural evolution,E2

where τ is the shear stress (from Cauchy stress measure), γ̇ is the shear rate (infinitesimal strain measure), λ is the structural parameter (a positive scalar quantity associated to the structural level of substance) and Gγ̇ is a functional form. In large number of works [5, 1821], Eq. (1) is represented by a linear viscoelastic model/mechanism (Maxwell, Jeffreys, Kelvin-Voigt, etc.) or combinations of that (in parallel or in series). However, it exists a problem with these approaches, they do not consider, the time/micro-structural dependence of the shear modulus (G) and viscosity (ημ and ην) in physical principles from which the constitutive equation system is obtained. Differently, the approach presented in [8] (the modified Jeffreys model) does not present this mistake, considering time/micro-structural evolution in the constitutive system development, and in this way, showing more coherence to represent thixotropic substances behavior.

Eqs. (1) and (2) connect micro-structural characteristic and viscoelasticity nature of the thixotropy. In the specialized literature, a considerable number of micro-structure types [2229] can be found, although only one single micro-structure type is considered in the development of the modified Jeffreys model. In this way, it is natural to imagine that it can exist more than one micro-structure type in a same thixotropic substance. In this chapter, a new perspective in terms of constitutive approach for thixotropic substances, with apparent-yield-stress nature and based on the modified Jeffreys model [8], is presented in terms of two different types of micro-structure (i.e., two structural parameters), and some of their properties are investigated in a rheological and thermodynamic sense. In the presented approach, each micro-structure is associated to a specific mechanism considered in the model (Maxwell-like element in parallel with a pure viscous Newtonian-like element), that is, one micro-structure type is associated to the Maxwellian element and another one to the pure viscous element. In this way, it is important to stand out that each of them (mechanisms) can react, under the shear loads, in distinct (but integrated) forms one from the other.

Aiming to extend the modified Jeffreys model for two distinguished micro-structures, in the next sections, the constitutive model is formally presented in terms of equations, the analysis of its thermodynamics consistence is done, points related to the characterization of the transition region are discussed and some illustrative numerical simulations of rheological tests are presented as well.

Advertisement

2. Main ideas on two types of micro-structures approach

The approach proposed in this work is based on a Jeffreys model [8] for the thixotropic system representation. In this case, a sketch of the model (Maxwell element in parallel with viscous element) is presented in Figure 1.

Figure 1.

Sketch of the two micro-structured thixotropic constitutive model.

It is important to point out the presence of two micro-structure types (λμ and λν) that are intimately related to the behavior of the shear modulus G=Gλν, and viscosity coefficients ην=ηνλν and ημ=ημλμ. In this way, under a certain shear load (τ), for the Maxwell element (.ν), one has

τν=ηνγ̇νv;E3
τ̇ν=Gγ̇νe+Ġγνe,E4

where γ̇νv is the viscous strain rate and γ̇νe is the elastic strain rate, respectively, and the total strain rate for the Maxwell element is γ̇ν=γ̇νv+γ̇νe. After some algebraic manipulation, it follows:

γ̇ν=τνην+τ̇νĠγνeG;
ηνγ̇ν=ηνGτ̇ν+1ηνĠG2τν.E5

From the viscous element (.μ),

γ̇μ=τμημτ̇μ=η̇μγ̇μ+ημγ¨μ,E6

and remembering that γ̇=γ̇ν=γ̇μ and τ=τν+τμτ̇=τ̇ν+τ̇μ, one has

τ̇=Gγ̇Gην1ηνĠG2τν+η̇μγ̇+ημγ¨;E7
ηνGτ̇+1ηνĠG2τ=ην+1ηνĠG2ημ+ηνη̇μGγ̇+ηνημGγ¨.E8

It is important to stand out that the time variation of the shear modulus and the viscosity coefficient are taken into account, which is in agreement with the expected physical behavior of thixotropic fluids [8, 30, 31].

It is important to keep in mind the fact that the parameters λμ (0λμ1) and λν (0λν1) are used to characterize the structural level of the substance. The structural parameter closer to 1 implies highly building-up state of micro-structure, and when it is closer to 0, the micro-structure is close to a fully broken-down state, in respective mechanism (.ν or .μ). In this sense, it is presented the two micro-structures version of constitutive model based on the work [8].

ηνGτ̇+1ηνĠG2τ=ην+1ηνĠG2ημ+ηνη̇μGγ̇+ηνημGγ¨;E9
dλνdt=1tνkν1λνβνKνλν6θγ¨+τνλνγ̇ςν;E10
dλμdt=1tμkμ1λμβμKμλμ6θγ¨+τμλμγ̇ςμ;E11
Gλν=Goλνexpmλν1;E12
ηνλν=ηoexpα1+α2λνexpα2λν;E13
ημλμ=ηoexpα3λμ.E14

In this case, there are the following positive parameters: , kν, kμ, βν, βμ, tν, tμ, Go, m, ςν, ςμ, Kν, Kμ, ηo, α1, α2 and α3. Some physical details related to these parameters can be found in [8, 32].

Advertisement

3. On the thermodynamic consistence

In this section, the thermodynamic consistence of constitutive model is analyzed in terms of the entropy-producing processes [33] associated with the concept of natural configuration [8, 3436]. The main objective is to verify the consistence of the constitutive equation to the Clausius-Duhem inequality.

Figure 2 presents a sketch of the configurations spaces used in this approach for the body B, where k̂oB is the reference configuration, k̂tB is the current/actual configuration, and k̂ntB is a family of natural configurations. In isothermal process, one is supposed that k̂ntB=k̂ntBλμλν, where λμ and λν are functions of the flow history. In this sense, it is assumed that the family of natural configurations can be parametrized by the structural level of the substance (λμ,λν). In this way, it has F¯¯k̂o (gradient of the motion from reference configuration to current configuration), F¯¯k̂nt (gradient of the motion from natural configuration to current configuration) and F¯¯k̂ok̂nt (gradient of the motion from reference configuration to natural configuration). In this way, denoting for each ξi-th relaxation mechanism.

ξiB¯¯k̂nt=ξiF¯¯k̂ntξiF¯¯k̂ntT;E15
ξiC¯¯k̂nt=ξiF¯¯k̂ntTξiF¯¯k̂nt;E16
ξiL¯¯k̂nt=ξiḞ¯¯k̂ok̂ntξiF¯¯k̂ok̂nt1;E17
ξiD¯¯k̂nt=symξiL¯¯k̂nt,E18

and the extensive thermodynamic Helmholtz potential (Ψ) can be represented as

Ψ=ΨλνI1νII1νIII1νIrννIIrννIIIrννλμI1μII1μIII1μIrμμIIrμμIIIrμμ;E19
Ψ̇=i=ν,μΨ,λiλ̇i+ξi=1riΨ,ξiB¯¯k̂nt:ξiḂ¯¯k̂nt,E20

where follows from chain rule Ψ¯¯,ξiB¯¯k̂nt=Ψ,Iξii+IξiiΨ,IIξiiI¯¯Ψ,IIξiiξiB¯¯k̂nt+IIIξiiΨ,IIIξiiξiB¯¯k̂ntT, with Iξii=trξiB¯¯k̂nt, IIξii=12trξiB¯¯k̂nt2trξiB¯¯k̂nt2 and IIIξi=detξiB¯¯k̂nt. Without loss of generality it is chosen k̂nt such that ξiF¯¯k̂nt=ξiV¯¯k̂nt (ξiC¯¯k̂nt=ξiB¯¯k̂nt), that is, for an appropriate rotated natural configuration k̂nt, it follows [8]:

I¯¯:ξiḂ¯¯k̂nt=2ξiB¯¯k̂nt:D¯¯ξiD¯¯k̂nt;E21
ξiB¯¯k̂nt:ξiḂ¯¯k̂nt=2ξiB¯¯k̂nt2:D¯¯ξiD¯¯k̂nt;E22
ξiB¯¯k̂ntT:ξiḂ¯¯k̂nt=2I¯¯:D¯¯ξiD¯¯k̂nt,E23

Figure 2.

Configuration spaces.

where D¯¯ is the symmetric part of velocity gradient (L¯¯). Returning to Eq. (20), one has

Ψ̇=i=ν,μΨ,λiλ̇i+ξi=1ri2Ψ,Iξi+IξiiΨ,IIξiiξiB¯¯k̂ntΨ,IIξiiξiB¯¯k̂nt2+IIIξiiΨ,IIIξiiI¯¯:D¯¯ξiD¯¯k̂nt.E24

Then, from the Clausius-Duhem inequality follows the specific rate of dissipation Ξ

Ξ=τ¯¯:D¯¯Ψ̇=τ¯¯i=ν,μξi=1ri2Ψ,Iξii+IξiiΨ,IIξiiξiB¯¯k̂ntΨ,IIξiiξiB¯¯k̂nt2+IIIξiiΨ,IIIξiiI¯¯:D¯¯i=ν,μΨ,λiλ̇i+i=ν,μξi=1ri2Ψ,Iξii+IξiiΨ,IIξiiξiB¯¯k̂ntΨ,IIξiiξiB¯¯k̂nt2+IIIξiiΨ,IIIξiiI¯¯:ξiD¯¯k̂nt,E25

where τ¯¯ is the Cauchy stress tensor.

This approach makes possible the analysis of the constitutive law and the nonnegativity of the entropy production in the context of irreversible thermodynamic processes. In this sense, it is assumed that total rate dissipation Ξ can be divided into three parts, each one associated to a specific process, as follows:

Ξk̂t=τ¯¯i=ν,μξi=1ri2Ψ,Iξii+IξiiΨ,IIξiiξiB¯¯k̂ntΨ,IIξiiξiB¯¯k̂nt2+IIIξiiΨ,IIIξiiI¯¯:D¯¯0;E26
Ξλ=i=ν,μΨ,λiλ̇i0;E27
Ξk̂nt=i=ν,μξi=1ri2Ψ,Iξii+IξiiΨ,IIξiiξiB¯¯k̂ntΨ,IIξiiξiB¯¯k̂nt2+IIIξiiΨ,IIIξiiI¯¯:ξiD¯¯k̂nt0,E28

where Ξk̂t is the dissipation rate due to changes in k̂tB), Ξλ is the dissipation rate due to changes in micro-structure and Ξk̂nt is the dissipation rate due to changes in k̂ntB.

Following similar steps to the work [8], one has

τ¯¯=i=ν,μξi=1riξiτ¯¯,E29

where

ξiτ¯¯=i=ν,μξi=1ri2Ψ,Iξii+IξiiΨ,IIξiiξiB¯¯k̂ntΨ,IIξiiξiB¯¯k̂nt2+IIIξiiΨ,IIIξiiI¯¯.E30

Thus, one has for each ξi-relaxation mechanism.

ηξiGξiξiτ¯¯+1ηξiĠ'ξiGξi2ξiτ¯¯=ηξiD¯¯,E31

where =̇L¯¯L¯¯T is the “Oldroyd time derivative”. Considering the two relaxation mechanisms, with rν=rμ=1ξντ¯¯=τ¯¯ν and ξμτ¯¯=τ¯¯μ, where ηξν=ην, Gξν=G, ηξμ=ημ and Gξμ+, one has.

ηνGτ¯¯ν+1ηνĠG2τ¯¯ν=ηνD¯¯;E32
τ¯¯μ=ημD¯¯.E33

Therefore, as τ¯¯=τ¯¯ν+τ¯¯μ one has

ηνGτ¯¯+1ηνĠG2τ¯¯=ην+1ηνĠG2ημ+ηνη̇μGD¯¯+ηνημGD¯¯.E34

Note that the isochoric motion constraint was not required in this analysis. In this way, associating each relaxation mechanism to a micro-structure, it is important to point out that depending on the nature (level of complexity) of the analyzed thixotropic substance, the reasoning presented here can be extrapolated for more than two relaxation mechanisms (i.e., more than two micro-structure types) aiming a consistent/coherent (in thermodynamic and rheological senses) approach of the model.

Advertisement

4. On the transition region

The objective of this topic is to analyze the transition/yielding region criteria under a theoretical and formal point of view. In this sense, it is important to stand out that the constitutive equation

ηνGτ¯¯+1ηνĠG2τ¯¯=ην+1ηνĠG2ημ+ηνη̇μGD¯¯+ηνημGD¯¯,E35

can be rewritten in a more appropriate form as follows:

θ1Π¯¯+Π¯¯=ηΓ̇¯¯+θ2Γ̇¯¯,E36

where

Π¯¯=τ¯¯G;E37
Γ̇¯¯=ημGD¯¯;E38
η=ην+ημημ;E39
θ1=ηνG;E40
θ2=ηνημGην+ημ.E41

It can be seen that the form of Eq. (36) is quite similar to the standard Jeffreys model. The quantities Π, Γ̇, η, θ1 and θ2 can be interpreted as a dimensionless stress, dimensionless strain, dimensionless apparent viscosity, relaxation time and retardation time, respectively. Note that since

G1¯·=G1,λνλ̇ν0;
ημG¯·=ημG,λμλ̇μ+ημG,λνλ̇ν0,

Eq. (36) corresponds to the standard Jeffreys model. In this way, it is clear the relation between the nonlinear viscoelastic region (associated to the thixotropic effect) and λ̇. Many works propose that an indicative for the beginning of the transition/yielding region is a region where occurs the modification in the behavior of the thixotropic substance, from the linear viscoelastic region to nonlinear viscoelastic region [3741] associating an specific strain value for that. However, this specific value is not constant [4145].

In this sense, taking into account the mapping in terms of variables in the actual configurations (x¯=xig¯i) and the respective displacements (u¯=uig¯i), it follows:

x1u1x1x2u2x1x3u3x1x1u1x2x2u2x2x3u3x2x1u1x3x2u2x3x3u3x3=1u1x1u2x1u3x1u1x21u2x2u3x2u1x3u2x31u3x3,

and in this context, one has the Jacobian (J˜)

J˜=det1u1x1u2x1u3x1u1x21u2x2u3x2u1x3u2x31u3x3,E42

and

J˜2=det12ε112ε122ε132ε2112ε222ε232ε312ε3212ε33;
=8ε11ε23ε32+ε12ε21ε33+ε13ε22ε31ε11ε22ε33ε12ε23ε31ε13ε32ε21
+4ε11ε33+ε11ε22+ε22ε33ε12ε21ε13ε31ε23ε32
2ε11+ε22+ε33+1;
=8JJJ+4JJ2J+1,E43

with

2εij=2ujxiu1xi2u2xi2u3xi2,ifi=j;ujxi+uixju1xiu1xju2xiu2xju3xiu3xj,ifij;

and

J=11!δjiεij;E44
JJ=12!δklijεjlεik;E45
JJJ=13!δrstijkεktεjsεir,E46

where δklij and δrstijk are the generalized Kronecker delta [46, 47]. Thus, the relationship between dV (differential of volume in the reference configuration) and dv (differential of volume in the actual configuration) can be described in the following way:

dV=8JJJ+4JJ2J+112dv.E47

Taking into account that the yield region can be characterized as a transition region. This region, in fact, can be treated as a singularity region. As the continuity axiom holds a suitable asymptotic behavior [4850], the concept of transition phenomenon has an asymptotic nature, and can be treated by limiting approaches. Physically, the transition region (transition/yielding) can be noted when a substance loses part of their original intrinsic properties, and from this point a new set of properties are raised in this substance, resulting in a new behavior. For example, in an elasto-plastic transition behavior, the transition state is related to a strain ellipsoid degeneracy, in a geometric sense, which can be transformed to an infinite cylinder, point or a pair of planes [51].

Considering the abovementioned lines, it can be imagined the nonlinear viscoelastic region as a direct mapping image from the linear viscoelastic region. Thus, the Jacobian (J˜), of this mapping, should translate the singularity on the transition region vicinity as an asymptotic extremely large deformation of the original microscopic element. In other words, taking into account Eq. (47), it can be seen that for a finite initial volume dV, with J˜0 (singularity on the neighbor of the transition region) implicates in dv, and in this way, it can write for the transition region the following condition:

8JJJ4JJ+2J=1,E48

or in an asymptotic sense on the transition region (t.r.)

limε¯¯t.r.8JJJ4JJ+2J=1.E49

It is important to stand out that this condition was stated in terms of the strain invariants, which can be rewritten in terms of the stress invariants or in terms of energy, by the constitutive relationships. Note the difficulty involved in expressing the strain tensor invariants in terms of the stress tensor invariants, due to the constitutive model’s form.

Remark 1. If it is taken the time rate ofEq. (48), one has

4JJJ¯·2JJ¯·+J̇=0;
4JJJε¯¯T:ε¯¯̇2JJtrε¯¯̇ε¯¯:ε¯¯̇+trε¯¯̇=0;
or
4JJJε¯¯T:D¯¯2JJtrD¯¯ε¯¯:D¯¯+trD¯¯=0.E50

Supposing the case that the strain rate is obtained as a response to known stress load, returning to Eq. (36), it follows:

Π¯¯+θ11Π¯¯=Gην+ημηνημΓ̇¯¯+Γ̇¯¯,E51

or

Γ̇¯¯+ϕΓ̇¯¯=ψ¯¯,E52

with

ψ¯¯=Π¯¯+θ11Π¯¯;
ϕ=Gην+ημημην.

Thus, one has

ϑΓ̇¯¯+ϑϕΓ̇¯¯=ϑΓ̇¯¯¯=ϑψ¯¯'Θ¯¯=ψ¯¯,E53

where

ϑt=expϕdt;
ψ¯¯t=ϑψ¯¯';
Θ¯¯=ϑΓ̇¯¯.

Note that

F¯¯F¯¯1Θ¯¯F¯¯TF¯¯T¯·=F¯¯˙F¯¯1Θ¯¯F¯¯TF¯¯T+F¯¯F¯¯1Θ¯¯F¯¯T¯·F¯¯T+F¯¯̇F¯¯1Θ¯¯F¯¯TF¯¯˙T;
Θ¯¯˙=L¯¯Θ¯¯+F¯¯F¯¯1Θ¯¯F¯¯T¯·F¯¯T+Θ¯¯L¯¯T;
Θ¯¯˙L¯¯Θ¯¯Θ¯¯L¯¯T=F¯¯F¯¯1Θ¯¯F¯¯T¯·F¯¯T,E54

and in this way, from Eq. (53), it follows:

Θ¯¯=F¯¯F¯¯1Θ¯¯F¯¯T¯·F¯¯T=ψ¯¯;
F¯¯1Θ¯¯F¯¯T¯·=F¯¯1ψ¯¯F¯¯T;
Θ¯¯t=F¯¯t0tF¯¯1tψ¯¯tF¯¯TtdtF¯¯Tt;
Γ̇¯¯t=1ϑtF¯¯t0tF¯¯1tψ¯¯tF¯¯TtdtF¯¯Tt,

or in other words

D¯¯t=GημϑtF¯¯t0tF¯¯1tψ¯¯tF¯¯TtdtF¯¯Tt;
=GημϑtF¯¯t0tF¯¯1tϑtΠ¯¯t+θ11Π¯¯tF¯¯TtdtF¯¯Tt;
D¯¯t=GημϑtF¯¯t0tF¯¯1texpGην+ημημηνdt
1G¯·τ¯¯t+G1τ¯¯,tt+τ¯¯tv¯tL¯¯tτ¯¯tτ¯¯tL¯¯Tt
+τ¯¯tηνF¯¯TtdtF¯¯Tt,E55

where L¯¯=v¯ (“Eulerian gradient”), and where v¯ is the velocity field. Consequently,

ε¯¯t=Gημ0t1ϑtF¯¯t0tF¯¯1texpGην+ημημηνt
G1τ¯¯,tt+τ¯¯tv¯tL¯¯tτ¯¯tτ¯¯tL¯¯Tt
+ην1τ¯¯tF¯¯TtdtF¯¯Ttdt,E56

since λμ and λν are taken as constants (pre-transition/yield region). In this way, the criteria Eq. (48) and/or Eq. (50) can be appropriately analyzed for a stress load excitation.

Advertisement

5. Numerical results

In this section, some illustrative numerical results are presented. Two types of rheological tests are analyzed, the constant shear rate test (CR) and the constant shear stress test (CS). In both cases it is necessary to define consistent initial conditions in relation to physical and mathematical principles. The finite difference method was implemented in MATLAB, with the following set of parameters: =1, η0=0.08Pa.s, α1=5, α2=α3=0.5, kνςν=103J/m3, kμςμ=102J/m3, βν=5, βμ=3, m=106, Kν=0.001Kgm1K1, Kμ=0.01Kgm1K1, G0=6.5Pa, tνςν=104s1 and tμςμ=103s1. It is important to point out that for both tests it was used Δt=103s. It is also important to comment that time-steps Δt=104s,105s and 106s were analyzed, but modifications in responses were not noted. The tolerance used for the Newton’s method was 106. Numerical results for real thixotropic substances (numerical/experimental comparative responses, regularization methods associated to nonlinear identification parameter problem, etc.) can be found in [52].

5.1. Constant shear rate test

In this section, it is considered the constant shear rate test, that is taken into account load conditions as γ̇t=Htγ̇0, where Ht represents the standard Heaviside function and γ̇0 it is a positive real constant. It is reasonable to think that in the begin of the test, the micro-structure is in a fully structured state (λμ0=λν0=1).

Figure 3 shows an interesting aspect, for the loads γ̇0=102s1 and 101s1 the behavior is close to a purely viscoelastic response. These behaviors are in agreement with Figures 4 and 5, for the same loads. Note that these two shear rates correspond to a low level of modification in λμ and λν. In these both cases, the stress of steady state is the maximum stress reached.

Figure 3.

τvs.t.

Figure 4.

λμvs.t.

Figure 5.

λνvs.t.

It is important to point out that for γ̇0=1s1 the thixotropic effect can ever be noted in stress response (Figure 3). The loads γ̇0=10s1 and 100s1 show a typical thixotropic behavior (Figure 3). These responses are related to considerable modifications in the structural nature of the substance, as can be seen in Figures 4 and 5. In these both cases, the maximum stress occurs before the steady state. It can be proved, in a theoretical/mathematical sense, that the point where the maximum breakdown rate occurs is before the stress peak. This fact is in agreement with expected behavior of thixotropic substance.

Other interesting result is related to the energy behavior of the thixotropic substance (Figures 6 and 7). Note the presence of a specific slope change in the region where higher rates of decrease on the micro-structures (λμ,λμ) levels occurred. The same behavior can be observed on experimental results. In fact, this can be explained, in a theoretical sense, via the following relationship

Power=kμςμ1λμβλμtμςμλ̇μλμ+kνςν1λνβλνtνςνλ̇νλν,E57

that comes from the rate Eqs. (11) and (12).

Figure 6.

Energyvs.t.

Figure 7.

Energyvs.γ.

5.2. Constant stress test

This chapter presents some points and results for the constant shear stress test. In this sense, it is considered loads as τt=Htτ0, where Ht represents the standard Heaviside function and τ0 is a positive real constant. It is assumed that in the begin of the test, the micro-structure is in a fully structured state (λμ0=λν0=1).

It can be seen in Figure 8, for the loads τ0=10, 20, 30, 40, 50 and 100Pa, a typical behavior for thixotropic substances under low level of constant stress loads. In this test, it can be seen the “Avalanche effect,” and their relation with the micro-structural level (λμ,λν) of the substance (Figures 9 and 10). In these cases, it can be observed the strain rate decrease, related to an increase of the construction parcel (cpμ) and decrease of the destruction parcel (dpμ) of λμ, as can be seen in Figures 11 and 12 respectively, where

cpμ=1tμkμ1λμβμ;E58
dpμ=Kμλμ6θγ¨+τμλμγ̇ςμ.E59

Figure 8.

γ̇vs.t.

Figure 9.

λμvs.t.

Figure 10.

λνvs.t.

Figure 11.

cpμvs.t.

Figure 12.

dpμvs.t.

For the Maxwell element (.ν), in the same region (the strain rate decreases), cpν is close to null value and dpν is increasing, but this increase level is not sufficient to stop the decreasing process of λ̇. A posterior increase of γ̇ is due to an abrupt decrease of linkages number (λμ,λν) before the steady state.

It is important to note the relationship between the structural nature and the behavior of the substance (Figures 9 and 10). Small changes in the values of λμ and λν were detected for stress loads 10 and 20Pa, in relation to the others. In these two cases γ̇ presents an nonincreasing behavior.

Other interesting result is presented in Figure 13, where it can be seen the energy behavior along the test. Note a changing on slope of the energy lines, tending to horizontal slope, in the γ̇ decreasing region and another one related to the γ̇ increasing region.

Figure 13.

Energyvs.t.

Advertisement

6. Concluding remarks

The main objective of this chapter is to investigate the constitutive model for thixotropic fluids based on [8] related to the existence of two different types of micro-structure, and their consistence in some rheological tests. The model and analysis presented here are based on well-established physics principles. In this sense, it is important to stand out some nonstandard points of the approach presented in this work for thixotropic modeling in respect to the others models. In this sense, it is important to stand out some points:

  • the shear modulus (Gλν), the viscosity coefficients (ημλμ and ηνλν) and their dependences from the two different micro-structure types are considered in the development of the constitutive model (Eqs. (9)(14));

  • the set of the rate equations (Eqs. (11)(12)) are related to well-established physical principles as the “reptation” model and the “Smoluchowski” theory of coagulation. It is important to point out that in the “Smoluchowski” theory of coagulation the effect of the Brownian motion is clearly taken into account;

  • the thermodynamic consistence of the model was analyzed (Section 3);

  • a theoretical criterium for the transition region based on the strain gradient mapping degeneration was discussed and exploited (Section 4), however it is important to stand out the necessity of more discussions and analysis on the relationships (Eqs. (48) and (56)) taking into account some additional characteristics (as temperature, …) and their consequences. These characteristics shall be considered in future work goals;

  • the illustrative numerical examples attest the capability of the model to predict the expected behavior of real thixotropic substances under some typical rheological tests (Section 5).

It is also important to comment that the developed ideas and presented in this work can be easily extended to approaches including more than two types of micro-structure (related to each specific relaxation mechanism) that are taken into account in the constitutive equation system. It is clear that the complexity level of the considered substance determines the necessity of these incorporations. In this context, it can be noted the versatility of the approach exposed here, presenting some interesting perspectives on thixotropic modeling that can be more explored in future works.

Advertisement

Symbology

τshear stress
γshear strain
Gshear modulus
ηdynamic viscosity
λstructural parameter
Gγ̇functional associated to structural parameter time evolution
δr…i…the generalized Kronecker delta
·total time derivative
··double total time derivative
Oldroyd time derivative
νassociated to the Maxwellian mechanism
μassociated to the pure viscous mechanism
¯first-order tensor
¯¯second-order tensor
inner product
:double inner product

References

  1. 1. Zhang X, Li W, Gong X. Thixotropy of MR shear-thickening fluids. Smart Materials and Structures. 2010;19(125012):1-6
  2. 2. El-Gendy H, Alcoutlabi M, Jemmett M, Deo M, Magda J, Venkatesan R, Montesi A. The propagation of pressure in a gelled waxy oil pipeline as studied by particle imaging velocimetry. AICHE Journal. 2012;58(3):302-311
  3. 3. Nguyen Q, Boger D. Thixotropic behaviour of concentrated bauxite residue suspensions. Rheologica Acta. 1985;24:427-437
  4. 4. Mujumdar A, Beris AN, Metzner AB. Transient phenomena in thixotropic systems. Journal of Non-Newtonian Fluid Mechanics. 2002;102:157-178
  5. 5. de Souza Mendes PR. Modeling the thixotropic behavior of structured fluids. Journal of Non-Newtonian Fluid Mechanics. 2009;164:66-75
  6. 6. Swift DL, Friedlander SK. The coagulation of hydrosols by Brownian motion and laminar shear flow. Journal of Colloid Science. 1964;19:621-647
  7. 7. Yaghouti MR, Rezakhanlou F, Hammond A. Coagulation, diffusion and the continuous Smoluchowski equation. Stochastic Processes and their Applications. 2009;119:3042-3080
  8. 8. Azikri de Deus HP, Negrão CRO, Franco AT. The modified Jeffreys model approach for elasto-viscoplastic thixotropic substances. Physics Letter A. 2016;380:585-595
  9. 9. Hammond A, Rezakhanlou F. Kinetic limit for a system of coagulating planar Brownian particles. Journal of Statistical Physics. 2006;124:997-1040
  10. 10. Niethammer B, Velázquez JJL. Self-similar solutions with fat tails for Smoluchowski’s coagulation equation with locally bounded kernels. Communications in Mathematical Physics. 2013;318:502-532
  11. 11. Pokrovskii VN. A justification of the reptation-tube dynamics of a linear macromolecule in the mesoscopic approach. Physica A. 2006;366:88-106
  12. 12. Öttinger HC. A thermodynamically admissible reptation model for fast flows of entangled polymers. Journal of Rheology. 1999;43:1461-1493
  13. 13. Mewis J. Thixotropy: A general review. Journal of Non-Newtonian Fluid Mechanics. 1979;6:1-20
  14. 14. Barnes H. Thixotropy: A review. Journal of Non-Newtonian Fluid Mechanics. 1997;70:1-33
  15. 15. Toorman EA. Modeling the thixotropic behaviour of dense cohesive sediment suspensions. Rheologica Acta. 1997;36:56-65
  16. 16. Ritter RA, Batycky JP. Numerical prediction of the pipeline flow characteristics of thixotropic liquids. SPE Journal. 1967;7:369-376
  17. 17. Dullaert K, Mewis J. A structural kinetics model for thixotropy. Journal of Non-Newtonian Fluid Mechanics. 2006;139:21-30
  18. 18. de Souza Mendes PR. Thixotropic elasto-viscoplastic model for structured fluids. Soft Matter. 2011;7:2471-2483
  19. 19. Azikri de Deus HP, Dupim GSP. Over strucutural nature of the thixotropic fluids behavior. Applied Mathematical Sciences. 2012;6:6871-6889
  20. 20. Azikri de Deus HP, Dupim GSP. On behavior of the thixotropic fluids. Physics Letters A. 2013;337:478-485
  21. 21. Azikri de Deus HP, Dupim GSP. Some aspects over thixotropic fluid behavior. Advanced Materials Research. 2013;629:623-634
  22. 22. Ardakani HA, Mitsoulis E, Hatzikiriakos SG. Thixotropic flow of toothpaste through extrusion dies. Journal of Non-Newtonian Fluid Mechanics. 2011;166:1262-1271
  23. 23. Oliveira GM, Rocha LLV, Franco AT, Negrão COR. Numerical simulation of the start-up of Bingham fluid flows in pipelines. Journal of Non-Newtonian Fluid Mechanics. 2010;165:1114-1128
  24. 24. Mewis J, Wagner NJ. Thixotropy. Advances in Colloid and Interface Science. 2009;147–148:214-227
  25. 25. Dullaert K, Mewis J. Thixotropy: Build-up and breakdown curves during flow. Journal of Rheology. 2005;49(6):1213-1230
  26. 26. Potanin A. 3D simulations of the flow of thixotropic fluids, in large-gap Couette and vane-cup geometries. Journal of Non-Newtonian Fluid Mechanics. 2010;165:299-312
  27. 27. Derksen JJ, Prashant NV. Simulations of complex flow of thixotropic liquids. Journal of Non-Newtonian Fluid Mechanics. 2009;160:65-75
  28. 28. Barnes HA. The yield stress – A review or ‘panta rei’ – Everything flows? Journal of Non-Newtonian Fluid Mechanics. 1999;81:133-178
  29. 29. Patil PD, Feng JJ, Hatzikiriakos SG. Constitutive modeling and flow simulation of polytetrafluoroethylene (PTFE) paste extrusion. Journal of Non-Newtonian Fluid Mechanics. 2006;139:44-53
  30. 30. Marrucci G. The free energy constitutive equation for polymer solutions from the dumbell model. Transactions. Society of Rheology. 1972;16:321-330
  31. 31. Marrucci G, Titomanlio G, Sarti GC. Testing of a constitutive equation for entangled networks by elongational and shear data of polymer melts. Rheologica Acta. 1973;12:269-275
  32. 32. Silva TABP. Aálise do Módulo de Cisalhamento Associado a Modelo de Jeffreys Modificado (in Portuguese). Master Dissertation. UTFPR, PPGEM, Brazil; 2017
  33. 33. Rajagopal KR, Srinivasa AR. On thermomechanical restrictions of continua. Proceedings of the Royal Society of London A. 2004;406:631-651
  34. 34. Rajagopal KR, Srinivasa AR. Inelastic behavior of materials. I. Theoretical underpinnings. International Journal of Plasticity. 1998;14:945-967
  35. 35. Rajagopal KR, Srinivasa AR. Inelastic behavior of materials II. Inelastic response. International Journal of Plasticity. 1998;14:967-995
  36. 36. Rajagopal KR, Srinivasa AR. A thermodynamic framework for rate-type fluid models. Journal of Non-Newtonian Fluid Mechanics. 2000;88:207-227
  37. 37. Marze S, Guillermic RM, Saint-James A. Oscillatory rheology of aqueous foams: Surfactant, liquid fraction, experimental protocol and aging effects. Soft Matter. 2009;5(9):1937
  38. 38. Mohan L, Pellet C, Cloitre M, Bonnecaze R. Local mobility and microstructure in periodically sheared soft particle glasses and their connection to macroscopic rheology. Journal of Rheology. 2013;57(3):1023
  39. 39. Mason TG, Bibette J, Weitz DA. Yielding and flow of monodisperse emulsions. Journal of Colloid and Interface Science. 1996;179(179):439-448
  40. 40. Souza Mendes PR, Thompson RL, Alicke AA, Leite RT. The quasilinear large-amplitude viscoelastic regime and its significance in the rheological characterization of soft matter. Journal of Rheology. 2014;58(2):537-561
  41. 41. Walls HJ, Caines SB, Sanchez AM, Khan SA. Yield stress and wall slip phenomena in colloidal silica gels. Journal of Rheology. 2003;47(4):847
  42. 42. Andrade DEV, Takii BA, Franco AT, Negrão CRO. The influence of the initial cooling condition on the flow curve of waxy crude oil. In: 15th Brazilian Congress of Thermal Sciences and Engineering. Belém: ABCM; 2014
  43. 43. Divoux T, Barentin C, Manneville S. Stress overshoot in a simple yield stress fluid: An extensive study combining rheology and velocimetry. Soft Matter. 2011;7:9335-9349
  44. 44. Fernandes RR, Andrade DEV, Franco AT, Negrão COR. Sampling methodology for rheological tests of drilling fluids: A study of the aging time and preshearing. In: 15th Brazilian Congress of Thermal Sciences and Engineering. Belém; 2014
  45. 45. Knauss WG, Zhu W. Nonlinearly viscoelastic behavior of polycarbonate. I response under pure shear. Mechanics of Time-Dependent Materials. 2002;6:231-269
  46. 46. Agarwa DC. Tensor Calculus and Riemannian Geometry. Meerut, India: Krishna Prakashan Media Ltd.; 2007
  47. 47. Lovelock D, Rund H. Tensors, Differential Forms, and Variational Principles. New York, USA: Dover Publications, INC.; 1989
  48. 48. Seth BR. Elastic-plastic transition in shells and tubes under pressure. Journal of Applied Mathematics and Mechanics. 1963;43:345-351
  49. 49. Seth BR. Generalized strain and transition concepts for elasti-plastic deformation, creep and relaxation. In: Proceedings of the Eleventh International Congress of Applied Mechanics; 1966. pp. 383–389
  50. 50. Seth BR. Transition Concept in Continuum Mechanics. Corvallis, USA: Seminar Lectures delivered in Oregon State University (unpublished); 1968
  51. 51. Purushothama CM. Elastic-plastic transition. Journal of Applied Mathematics and Mechanics. 45:401-408
  52. 52. Morinigo V Jr. Verificação Teórico-numéica de Modelo Constitutivo Aplicado a Fluidos Tixotrópicos Compostos por Duas Estruturas Distintas (in Portuguese). TCC, Federal University of Technology-Paraná – UTFPR, Brazil; 2017

Written By

Hilbeth P. Azikri de Deus and Mikhail Itskov

Submitted: 22 November 2017 Reviewed: 26 February 2018 Published: 05 November 2018