Explicit Solutions To Phase-Change Problems and Applications

2016 Computational Mathematics Annual Program Review
Arlington, VA (USA), August 8-11, 2016
EXPLICIT SOLUTIONS TO PHASE-CHANGE

PROBLEMS AND APPLICATIONS
Domingo Alberto Tarzia
Grant AFOSR – SOARD FA 9550-14-1-0122
CONICET - Depto. Matemática, FCE, Univ. Austral,

Paraguay 1950, S2000FZF Rosario, Argentina.
E-mail: DTarzia@austral.edu.ar
Key Words: Free boundary problems, Lamé-Clapeyron-Stefan problem, Phase-change problems, Heat-
Diffusion equation, Fractional diffusion, Caputo fractional derivative, Explicit solutions, Neumann
solution, Unknown thermal coefficients, Over-specified boundary condition, Mushy region, Convective
boundary condition.
Mathematics Subject Classification 2010: 26A33, 35C05, 35R11, 35R35, 80A22.

2
ABSTRACT
We obtain explicit solutions (in a closed form) for the following one-dimensional phase-
change problems:
• Two-phase solidification process with a convective boundary condition (T.
Thermal Sci (2016), In Press);
• One-phase solidification process with a mushy region and a convective boundary
condition (T., J. Appl. Math. (2015));
• One-phase solidification process with a mushy region and a heat flux boundary
condition (T., J. Appl. Math. (2015));
• Determination of one unknown thermal coefficient through a solidification problem
with an over-specified boundary condition (Ceretani-T., Math. Problems in Eng.
(2015));
AFOSR - Comput. Math., Arlington (VA, USA), 8-11 August 2016 Tarzia, Explicit Solutions to Phase-Change Problems & Appl.
3
• Simultaneous determination of two unknown thermal coefficients through a

solidification problem with an over-specified boundary condition (Ceretani-T., JP
Journal Heat Mass Transfer (2016));
• Two-phase fractional Stefan problem with a temperature boundary condition
(Roscani-T., Adv. Math. Sci. Appl. (2014));
• Two-phase fractional Stefan problem with a heat flux boundary condition (Roscani-
T., In progress);
• Determination of one unknown thermal coefficient through a one-phase fractional
Stefan problem with an over-specified boundary condition (T. Appl. Math. (2015));
4
INTRODUCTION
Heat and mass transfer problems with a phase-change process (known in the literature as
Stefan problem or Lamé-Clapeyron-Stefan problem) begins by modeling the solidification
of the Earth planet (1831). Phase-change process such as melting, freezing, solidification,
sublimation, desublimation, thawing and drying have been studied in the last century due to
their wide scientific and technological applications. Some of them are:
i) Continuous casting of steel (e.g., ingots, slabs);

ii) Binary alloy solidification problem (liquidus and solidus curves);
iii) Permafrost process (e.g., frozen ground in cold region of the Earth planet);
iv) Manufacture of glass and crystal;
v) Phase-change process with a mushy region;
vi) Thawing in a saturated porous medium with the influence of the pressure on the melting
temperature;
vii) Coupled heat and mass transfer for a phase-change process (e.g., Luikov systems);
viii) Coupled heat and mass transfer during the freezing of the high-water content materials
with two free boundaries: the freezing and sublimation fronts (e.g., for food
preservation);
ix) Ablation process (e.g., entry of a spacecraft to the atmosphere of the Earth planet);
x) Welding process;
5
xi) Solidification of a supercooled liquid (temperature of the liquid is below of the melting
temperature);
xii) Phase-change materials to store solar energy as the latent heat of melting (e.g., heating of
a house);
and some related free boundary problems as:
xiii) Oxygen diffusion-consumption in a tissue (equivalent to the phase-change problem);

xiv) Isothermal diffusion-reaction process of a gas with a solid;
xv) Gas flow through a porous medium;
xvi) Penetration of solvent in polymers;
xvii) Filtration of water through oil in a porous medium;
xviii) Nutrient uptake by a growing root system.
6
A review of a long bibliography on moving and free boundary problems for phase-
change materials (PCM) for the heat equation is given in:
T., "A bibliography on moving-free boundary problems for the heat-diffusion equation. The
Stefan and related problems", MAT - Serie A, 2 (2000), 1-297 (with 5869 titles). Available
from:
http://web.austral.edu.ar/descargas/facultad-cienciasEmpresariales/mat/Tarzia-MAT-
SerieA-2(2000).pdf
Obtaining explicit solution we have:

i) a tool to get super and sub-solution for general conditions;
ii) a benchmark solution for checking on the correctness of a computer code
A review on explicit solutions for heat and mass transfer problems is given in:
T., “Explicit and Approximated Solutions for Heat and Mass Transfer Problems with a
Moving Interface”, Chapter 20, In Advanced Topics in Mass Transfer, Mohamed El-Amin
(Ed.), InTech Open Access Publisher, Rijeka (2011), pp. 439-484. Available from:
http://www.intechopen.com/articles/show/title/explicit-and-approximated-solutions-for-heat-
and-mass-transfer-problems-with-a-moving-interface
7
OUTLINE
I. Classical two-phase Stefan problems
 Classical Neumann solution for a temperature boundary condition at the fixed face x = 0
 Classical Neumann solution with a heat flux boundary condition at the fixed face x = 0
 Classical Neumann solution with a convective boundary condition at the fixed face x = 0
II. Classical one-phase Stefan problem with a simple mushy region

 One-phase with a convective boundary condition
 One-phase with a heat flux boundary condition
 Applications:
o Determination of one unknown thermal coefficient
o Simultaneous determination of two unknown thermal coefficients
8
III. Phase-change problems for the fractional heat-diffusion equation

 Fractional derivatives
o Caputo fractional derivative
o Wright and Mainardi Functions
o Basic properties
 Two-phases fractional Stefan problems:

o Generalized Neumann solution with a temperature boundary condition
o Generalized Neumann solution with a heat flux boundary condition
 One-phase fractional Stefan problem:

IV. Papers which are not reviewed
9
I. Classical two-phase Stefan problems
 Classical Neumann solution with a temperature boundary condition at the fixed face x = 0
(Weber, Book (1901)).
 Classical Neumann solution with a heat flux boundary condition at the fixed face x = 0
(T., Quart. Appl. Math. (1981)).
 Classical Neumann solution with a convective boundary condition at the fixed face x = 0
(T. Thermal Science (2016), In Press)
10
I.1) CLASSICAL NEUMANN SOLUTION (MELTING PROCESS) WITH A

TEMPERATURE BOUNDARY CONDITION
Problem ( P1 ) The two-phase Lamé-Clapeyron-Stefan problem is given by: find the free
boundary x = s(t ) , and the temperature T = T ( x , t ) defined by:
T ( x , t ) > T f if 0 < x < s(t ), t > 0

T ( x , t ) = T f if x = s(t ), t > 0 (T1)

Ts ( x , t ) < T f if s(t ) < x , t > 0
such that the following equations and conditions are satisfied:
 csTst − ksTsxx = 0, x > s(t ), t > 0 , (T2)
 cT t − kT xx = 0, 0 < x < s(t ), t > 0 , (T3)
s(0) = 0 , (T4)
Ts ( x ,0) = Ts ( +∞ , t ) = Ti < T f , x > 0, t > 0 , (T5)
Ts (s(t ), t ) = T f , t > 0, (T6)
Tl (s(t ), t ) = T f , t > 0, (T7)
ksTsx ( s(t ), t ) − kT x ( s(t ), t ) =  s(t ), t > 0 , (T8)
T (0, t ) = T0 > T f , t > 0. (T9)
11
12
Theorem T1 [Weber, Book (1901)] Let Ti < T f < T0 be. The solution of the problem ( P1 ) is
the classical Neumann explicit solution given by:
T0 − T f  x 
T ( x , t ) = T0 − erf   , 0 ≤ x ≤ s(t ), t > 0 , (T10)
erf ( ) 
 2  t 
T f − Ti  x 
Ts ( x , t ) = Ti + erfc   , s(t ) ≤ x , t > 0 , (T11)
erfc( ) 
 2  st 
 k ks s 
s(t ) = 2  st ,  = =  ,  s =
2
= s ,  =
2
, (T12)
  c  cs  
where the dimensionless coefficient  > 0 is the unique solution of the following equation:
G( x ) = x , x > 0 (T13)
with
13
G( x ) = b4 G2 ( x ) − b3 G1 ( x ) , (T14)
x
2
∫
2
erf ( x ) = exp( − u )du , erfc( x ) = 1 − erf ( x ) , (T15)
 0
exp( −x 2 ) exp( −x 2 )
G1 ( x ) = , G2 ( x ) = , (T16)
erfc( x ) erf ( x )
c s (T f − Ti ) k (T0 − T f )
b3 = > 0, b4 = > 0. (T17)
     s 
14
I.2) CLASSICAL NEUMANN SOLUTION (MELTING PROCESS) WITH A HEAT

FLUX BOUNDARY CONDITION
Problem ( P2 ) Find the free boundary x = s(t ) , and the temperature T = T ( x , t ) defined by:
T ( x , t ) > T f if 0 < x < s(t ), t > 0

T ( x , t ) = T f if x = s(t ), t > 0 (F1)

Ts ( x , t ) < T f if s(t ) < x , t > 0
 csTs − ksTs = 0,
t xx
x > s(t ), t > 0 , (F2)
 cT − kT = 0,
t xx
0 < x < s(t ), t > 0 , (F3)
s(0) = 0 , (F4)
Ts ( x ,0) = Ts ( +∞ , t ) = Ti < T f , x > 0, t > 0 , (F5)
Ts (s(t ), t ) = T f , t > 0, (F6)
Tl (s(t ), t ) = T f , t > 0, (F7)
ksTsx ( s(t ), t ) − kT x ( s(t ), t ) =  s(t ), t > 0 , (F8)
q
kT x (0, t ) = − 10 , t > 0. (F9)
t 2
15
16
Theorem F1 [T., Quart. Appl. Math. (1981)] Let Ti < T f be.
a) If the coefficient q0 satisfies the inequality:
ks (T f − Ti )
q0 > , (F10)
 s
then there exists an instantaneous phase-change (melting) process and the problem ( P2 ) has the
unique Neumann explicit solution of a similarity type given by:
q0     x 
T ( x , t ) = T f +  erf ( F ) − erf    , 0 ≤ x ≤ s(t ), t > 0 (F11)
k   2  t  
T f − Ti  x 
Ts ( x , t ) = Ti + erfc   , s(t ) ≤ x , t > 0 , (F12)
erfc( F ) 
 2  st 
 k ks 
s(t ) = 2 F  st , 
  = ,  s = , (F13)
  c   cs 
17
where the dimensionless coefficient  F > 0 is the unique solution of the following equation:
GF ( x ) = x , x > 0 (F14)
with
GF ( x ) = b5 exp( − 2 x 2 ) − b3 G1 ( x ) , (F15)
s c s (T f − Ti ) q0
 =2
> 0, b3 = > 0, b5 = > 0. (F16)
    s
b) If the coefficient q0 satisfies the inequalities:
ks (T f − Ti )
0 < q0 ≤ , (F17)
 s
then the problem ( P2 ) is a classical heat transfer problem for the initial solid phase whose
solution is given by:
q0  s  x 
Ts ( x , t ) = Ti + erfc   , x > 0, t > 0 . (F18)
ks 
 2  st 
18
I.3) CLASSICAL NEUMANN SOLUTION (SOLIDIFICATION PROCESS) WITH A

CONVECTIVE BOUNDARY CONDITION
Problem ( P3 ) Find the free boundary x = s(t ) , and the temperature T = T ( x , t ) defined by:
Ts ( x , t ) < T f if 0 < x < s(t ), t > 0

T ( x , t ) = T f if x = s(t ), t > 0 (C1)

T ( x , t ) > T f if s(t ) < x , t > 0
 csTs − ksTs = 0,
t xx
0 < x < s(t ), t > 0 , (C2)
 cT − kT = 0,
t xx
x > s(t ), t > 0 , (C3)
s(0) = 0 , (C4)
T ( x ,0) = T ( +∞ , t ) = Ti > T f , x > 0, t > 0 , (C5)
Ts (s(t ), t ) = T f , t > 0, (C6)
Tl (s(t ), t ) = T f , t > 0, (C7)
ksTsx ( s(t ), t ) − kT x ( s(t ), t ) =  s(t ), t > 0 , (C8)
−1 2
ksTsx (0, t ) = h0 t (Ts (0, t ) − T∞ ) , t > 0 ( h0 > 0) . (C9)
19
Theorem 1 Let T∞ < T f < Ti be.
a) If the coefficient h0 satisfies the inequality:
kl Ti − T f
h0 > , (C10)
  T f − T∞
then there exists an instantaneous solidification process and the free boundary problem ( P3 )
has the unique solution of a similarity type given by:
 h0  s  x 
(T f − T∞ )  1 + erf 
 k  2  t  
 s  s 
Ts ( x , t ) = T∞ +
h0  s   
1+ erf   
ks   s 
    x 
erf    − erf  
h0  s (T f − T∞ )  s  2  st 
= Tf − (C11)
ks h0  s   
1+ erf   
ks  s 
20
 x    x 
erfc    erfc  
 2  t  = T + (T − T )  1 −  2  t   ,
T ( x , t ) = Ti − (Ti − T f ) f  (C12)
erfc( ) 
f i
erfc( )
 
 
 
s(t ) = 2 C  t , (C13)
and the dimensionless coefficient C > 0 satisfies the following equation:
F( x ) = x , x > 0 , (C14)
where the function F and the parameters bi are given by:
exp( −bx 2 )
F( x ) = b1 − b3 F1 ( x ) , (C15)
1 + b2 erf ( x b )
h0 (T f − T∞ ) h0
b1 = > 0, b2 =  s > 0 . (C16)
  ks
21
b) If the coefficient h0 satisfies the following inequalities:

kl Ti − T f
0 < h0 ≤ . (C17)
  T f − T∞
then the free boundary problem ( P3 ) is a classical heat transfer problem for the initial liquid
phase whose solution is given by:
kTi  x 
T∞ + + (Ti − T∞ )erf  
h0    2  t
  Ti − T∞  k  x 
T ( x , t ) = = T∞ +  + erf 
k k  h0    2  t  
1+ 1+    
h0   h0  
Ti − T∞  x 
= Ti − erfc  , x > 0, t > 0 .
 2  t 
(C18)
k
1+   
h0  
22
II. Classical one-phase Stefan problem with a simple mushy region
 One-phase with a convective boundary condition (T., J. Appl. Math. (2015))
 One-phase with a heat flux boundary condition (T., J. Appl. Math. (2015))
 Applications:
o Determination of one unknown thermal coefficient (Ceretani-T., Math. Problems

in Eng. (2015))
o Simultaneous determination of two unknown thermal coefficients (Ceretani-T., JP

Journal Heat Mass Transfer (2016))
23
II.1 Classical one-phase Stefan problem with a simple mushy region and a convective
boundary condition at the fixed face x = 0
Solidification process with a convective boundary condition:

It consists in finding the free boundaries x = s(t ) and x = r (t ) , and the temperature
T = T ( x , t ) such that the following conditions are verified:
 cTt − kTxx = 0 , 0 < x < s ( t ) , t > 0 ( = k /  c ) (SC1)
T ( s (t ) , t ) = 0 , t>0 (SC2)
kTx (s(t ), t ) =  [ s(t ) + (1 −  )r(t )], t > 0 (SC3)
Tx (s(t ), t )(r (t ) − s(t )) =  > 0, t > 0 (with  > 0) (SC4)
s(0) = r (0) = 0 (SC5)
h0
kTx ( 0, t ) = (T (0, t ) + D∞ ) , t > 0 (with h0 > 0, D∞ > 0) . (SC6)
t
24
Theorem.
If the coefficient h0 satisfies the inequality
1  (1 −  ) k
h0 > (SC7)
D∞ 2
then the solution to the problem (SC1)-(SC6) is given by:
h0 D∞    x 
erf ( )  erf  
T (x , t ) = − k 1 −  2 t , 0 < x < s(t ), t > 0 , (SC8)
h  erf ( )
1+ 0 erf ( )  

k 
s(t ) = 2  t , t > 0 , (SC9)
r (t ) = 2   t , t > 0 , (SC10)
k 2 h0  
 = + e 1 + erf ( ) , (SC11)
2 D∞ h0   k 
25
where the coefficient  is given as the unique solution of the equation:
D∞c
F3 ( x ) = G( x ), x > 0 , (SC12)
 
with
exp( −x 2 )  (1 −  )  1
F3 ( x ) = , G( x ) = x + , x>0 (SC13)
k 2D∞ F3 ( x )
+ erf ( x )
h0  .
26
II.2 Classical one-phase Stefan problem with a simple mushy region and a heat flux
boundary condition at the fixed face x = 0
Solidification process with a heat flux condition:

It consists in finding the free boundaries x = s(t ) and x = r (t ) , and the temperature
T = T ( x , t ) such that the following conditions are verified:
 cTt − kTxx = 0 , 0 < x < s ( t ) , t > 0 ( = k /  c ) (SF1)
T ( s (t ) , t ) = 0 , t>0 (SF2)
kTx (s(t ), t ) =  [ s(t ) + (1 −  )r(t )], t > 0 (SF3)
Tx (s(t ), t )(r (t ) − s(t )) =  > 0, t > 0 (with  > 0) (SF4)
s(0) = r (0) = 0 (SF5)

q0
kTx ( 0, t ) = , t > 0 (with q 0 > 0) . (SF6)
t
27
Theorem.
If the coefficient q 0 satisfies the inequality
 (1 −  ) k
q0 > (SF7)
2
then the solution to the problem (SF1)-(SF6) is given by:
  x 
erf  
q0  erf ( )   2 t ,
T (x , t ) = − 1 − 0 < x < s(t ), t > 0 , (SF8)
k  erf ( ) 
 
s(t ) = 2  t , t > 0 , (SF9)
r (t ) = 2  t , t > 0 , (SF10)
k 2
 = + e , (SF11)
2 q0 
28
where the coefficient  is given as the unique solution of the equation:
q0
G3 ( x ) = , x > 0, (SF12)
 
with
  (1 −  )k x2  x2
G3 ( x ) =  x + e e , x > 0 (SF13)
 2 q0  
.
29
III. Determination of unknown thermal coefficients through a one-phase Stefan problem

with a simple mushy region and an over-specified condition at the fixed face x = 0
III.1 Determination of one unknown thermal coefficient
The problem consists in finding the free boundaries x = s(t ) and x = r(t ) , the temperature
T = T ( x , t ) and one unknown termal coefficient among {k , c,  , ,  ,  } such that the following
conditions must be verified:
 cTt − kTxx = 0 , 0 < x < s ( t ) , t > 0 ( = k /  c ) (1FC1)
T ( s (t ) , t ) = 0 , t>0 (1FC2)
kTx (s(t ), t ) =  [ s(t ) + (1 −  )r(t )], t > 0 (1FC3)
Tx (s(t ), t )(r (t ) − s(t )) =  , t > 0 (1FC4)

s(0) = r (0) = 0 (1FC5)
q
kTx ( 0, t ) = 0 , t > 0 (with q0 > 0) (1FC6)
t
h0
kTx ( 0, t ) = (T (0, t ) + D∞ ) , t > 0 (with h0 > 0, D∞ > 0) . (1FC7)
t
30
The problem can be divided in 6 different cases:
Case # One Unknown Coefficient
1 Latent heat of fusion by unit of mass: 
2 First parameter of the mushy region: 
3 Second parameter of the mushy region: 
4 Thermal conductivity: k
5 Density of mass: 
6 Specific heat by unit of mass: c
31
Theorem. If h0 and q0 are two positive numbers determinated experimentally then the
solution of the 6 cases are given by:
q0    x  
T (x , t ) =  erf  2  t  − erf ( ) , 0 < x < s(t ), t > 0 , (1FC8)
k    
s (t ) = 2  t , t > 0 , (1FC9)
r (t ) = 2   t , t > 0 , (1FC10)
k 2
 = + e . (1FC11)
2 q0 
and the coefficient  and the unknown thermal coefficient are given by the following Table 1.
Table 1: Formulae for the problem with a convective over-specified boundary condition. Explicit formulae for the unknown thermal coefficient l,
γ, , k, ρ or c and the coefficient ξ (or the equation that it must satisfy) and the corresponding restrictions on data that guarantee their validity.
Case Thermal coefficient Coefficient ξ that characterizes Restrictions

the free boundary x = s(t) on data
q0 exp (−ξ 2 )
q
c h
q
1 l= ρk ξ+ γ(1−)
√
kρc
i ξ = erf −1 D∞
q0
kρc
π 1− q0
h0 D∞ (R1), (R2)
exp (ξ 2 )
2q0
q q
2 γ= 2q√
0
(1−) kρc
q0
l
c
ρk − ξ exp (ξ 2 ) exp (−2ξ 2 ) ξ = erf −1 D∞
q0
kρc
π 1− q0
h0 D∞ (R1), (R2)
(R3)
q q
3 =1− 2q0
√
γ kρc
q0
l
c
ρk − ξ exp (ξ 2 ) exp (−2ξ 2 ) ξ = erf −1 D∞
q0
kρc
π 1− q0
h0 D∞ (R1), (R2)
(R3), (R4)
2
π q0 erf(ξ)
4 k= ρc q
Unique positive solution (R1)
D∞ 1− h D0∞
0
of the equation (E4)
2
π q0 erf(ξ)
5 ρ= kc q
Unique positive solution (R1)
D∞ 1− h D0∞
0
2
π q0 erf(ξ)
6 c= ρk q
Unique positive solution (R1), (R5)
D∞ 1− h D0∞
0
Restrictions on data
q0
1− >0 (R1)
h0 D ∞
r
D∞ kρc q0
1− <1 (R2)
q0 π h0 D ∞
r !! r
−1 D ∞ kρc q 0 q0 c
f2 erf 1− < (R3)
q0 π h0 D ∞ l ρk
√
 " 
r !#2
γ kρc D ∞ kρc q 0
exp 2 erf −1 1− +
2q0 q0 π h0 D∞
(R4)
r !! r
−1 D ∞ kρc q 0 q0 c
f2 erf 1− >
q0 π h0 D ∞ l ρk
2
q0 1 2q0
1− < − γ(1 − ) (R5)
h0 D∞ D∞ ρlk
Definitions of functions (x > 0)
f2 (x) = x exp (x2 )

√
 
πγ(1 − )
f4 (x) = x + erf(x) exp (x2 ) erf(x) exp (x2 )
q0
2D∞ 1 − h0 D∞
√
 
x πγ(1 − )
f6 (x) =  + exp (x2 ) exp (x2 )
erf(x) 2D 1− q0
∞ h0 D∞
Ecuations for ξ

cD∞ q0
f4 (x) = √ 1− x>0 (E4)
πl h0 D∞
√ 2
πq
f6 (x) = 0 x>0 (E6)
ρlkD∞ 1 − h0qD0 ∞
32
III.2 Simultaneous Determination of two thermal coefficients
We follow the previous section III.1 and we consider that the left boundary of the mushy
region is a moving boundary given by the expression:
s (t ) = 2 t , t > 0 , (2FC1)
where  > 0 is a positive parameter obtained experimentally.
33
The problem consists in finding the free boundary x = r(t ) , the temperature T = T (x , t ) and
two unknown termal coefficients among {k , c,  , ,  ,  } such that the following conditions are
verified:
 cTt − kTxx = 0 , 0 < x < s ( t ) , t > 0 ( = k /  c ) (2FC2)
T ( s (t ) , t ) = 0 , t>0 (2FC3)
kTx (s(t ), t ) =  [ s(t ) + (1 −  )r(t )], t > 0 (2FC4)
Tx (s(t ), t )(r (t ) − s(t )) =  , t > 0 (2FC5)
s(0) = r (0) = 0 (2FC6)

q0
kTx ( 0, t ) = , t > 0 (with q0 > 0) (2FC7)
t
h
kTx ( 0, t ) = 0 (T (0, t ) + D∞ ) , t > 0 (with h0 > 0, D∞ > 0) . (2FC8)
t
34
The problem can be divided in 15 different cases:
Case # Two Unknown Coefficients

1  and 
2  and 
3  and 
4  and k
5  and 
6  and c
7  and k
8  and 
9  and c
10  and k
11  and 
12  and c
13 k and 
14 k and c
15  and c
35
Theorem. If h0 , q0 and  are three positive numbers determinated experimentally then the
solution of the 15 cases are given by:
q0    x  
T (x , t ) =  erf  2  t  − erf ( ) , 0 < x < s(t ), t > 0 , (2FC9)
k    
  
2
 k e  
r (t ) =  2 +  t , t > 0, (2FC10)
 q0 
 
and the two unknown thermal coefficients are given by the following Table 2.
Table 2: Formulae for the problem with a convective over-specified boundary condition. Explicit formulae for the two unknown thermal coefficients
chosen among l, γ, , k, ρ or c, the equation for the coefficient ξ and the corresponding restrictions on data that guarantee their validity.
Case Thermal coefficients Equation for

q Restrictions
ξ = √σα = σ ρc k on data
1 0<<1 – (R0), (R1)

2q0 σ q0
γ= k(1−) ρlσ exp (−σ 2 /d) − 1 exp (−σ 2 /d)
2 0<<1 – (R0)
2
l= h q0 exp (−σ /d) i
γk(1−)
ρσ 1+ 2q σ exp (σ 2 /d)
0
3 γ>0 – (R0)
2
l= h q0 exp (−σ /d) i
γk(1−)
ρσ 1+ 2q σ exp (σ 2 /d)
0
4 = 1 − F4 (ξ) (E4) (R2), (R3), (R4),

2
k = ρc σξ (R5), (R6) or:
(R2), (R3), (R4),
(R7), (R8) or:
(R2), (R3), (R4),
(R9)
5 = 1 − F5 (ξ) (E5) (R12)

2
ρ = kc σξ
6 = 1 − F6 (ξ) (E5) (R13), (R14), (R15)

2
c = kρ σξ or:
(R13), (R16), (R17)

2q0 q0
7 γ= σρc(1−) ρlσ exp (−ξ 2 ) − 1 ξ 2 exp (−ξ 2 ) (E4) (R2), (R3), (R4)
2
k = ρc σξ
Case Thermal coefficients Equation q
for Restrictions
ξ = √d = σ ρc
σ
k on data

2q0 σ q0 cσ exp (−ξ 2 )
8 γ= k(1−) lk ξ − 1 exp (−ξ 2 ) (E5) (R2), (R17), (R18)
2
k ξ
ρ= c σ

2q0 σ q0
9 γ= k(1−) ρlσ exp (−ξ 2 ) − 1 exp (−ξ 2 ) (E5) (R3), (R13), (R17)
2
c = kρ σξ

q0 1 exp (ξ 2 )
10 l= ρσ ρcσγ(1−) ξ2
exp (−ξ 2 ) (E4) (R2)
1+ 2q0
2
σ
k = ρc ξ

q0 cσ 1 exp (−ξ 2 )
11 l= k γk(1−) ξ (E5) (R2), (R17)
1+ 2q σ exp (ξ 2 )
0 2
k ξ
ρ= c σ

q0 1
12 l= ρσ γk(1−) exp (−ξ 2 ) (E5) (R2), (R17)
1+ 2q0 σ
exp (ξ 2 )
2
k ξ
c= ρ σ
√
13 k= πq0 σ G5 (ξ) (E13) (R2)
q
D∞ 1− h D0
√ 0 ∞
πq0
ρ = cσD∞
G4 (ξ)
√
14 k= πq0 σ G5 (ξ) (E14) (R2), (R3), (R19)
q
D∞ 1− h D0
√ 0 ∞
c= πq0 G4 (ξ)
q
σρD∞ 1− h D0
0 ∞
q0 exp (−ξ 2 )
15 ρ= lσ 1+ γk(1−) exp (ξ 2 ) (E5) (R2), (R17)
h 2q σ i0
c= kl
σq0 1 + γk(1−)
2q0 σ exp (ξ 2 ) ξ 2 exp (ξ 2 )
Restrictions on data

σ kD∞ q0
erf√ = √ 1− (R0)
α q0 πα h0 D ∞
q0
0< exp (−σ 2 /d) − 1 (R1)
ρlσ
q0
0<1− (R2)
h0 D ∞
q0
0< −1 (R3)
ρlσ
√ s !
q0 πq0 q0
1− < g4 ln (R4)
h0 D ∞ σρcD∞ ρlσ
F4 (η) > 1 (R5)
where η is the only one positive solution to the equation:
q0
(1 − 2x2 ) = (1 − x2 ) exp(x2 ), (1)
ρlσ
√ √
q0 πq0 q0 πq0
1− < G4 (ζ1 ) o 1− > G4 (ζ2 ) (R6)
h0 D∞ σρcD∞ h0 D∞ σρcD∞
where ζ1 and ζ2 are the only two positive solutions to the equation:
F4 (x) = 1,
F4 (η) = 1 (R7)
where η is the only one positive solution to the equation (1).
√
q0 πq0
1− 6= G4 (η) (R8)
h0 D∞ σρcD∞
F4 (η) < 1 (R9)
√ s ! √ s !
πq0 1 q0 πq0 q0
G4 ln <1− < G4 ln (R10)
σρcD∞ ν4 h0 D∞ σρcD∞ ρlσ
where:  
v
ρlσ q0 u 2γc 
ν4 = ln 1 + t1 +
 u 
2q0 ρlσ q0
l ln ρlσ

2q0 q0 q0
0< ln −1 −1 (R11)
ργcσ ρlσ ρlσ
√ √
πq0 σ q0 2q0 σ πq0 σ
G5 (ζ1 ) < 1 − < min , G5 (ζ2 ) (R12)
kD∞ h0 D ∞ kD∞ kD∞
where ζ1 and ζ2 are the only one positive solutions to the equations:
q0 σc
exp (−x2 ) = x2
lk
y:

q0 σc 2 γk
exp (−x ) = exp (x ) + 1 x2 ,
2
lk 2q0 σ
respectively.
√ s !
πq0 σ q0 q0
G5 ln <1− (R13)
kD∞ ρlσ h0 D ∞
q0 γk
≥ +1 (R14)
ρlσ 2q0 σ
( √ s !)
q0 2q0 σ πq0 σ 1
1− < min , G5 ln (R15)
h0 D ∞ kD∞ kD∞ ν6
where: " s #
ρlσ 2γk
ν6 = 1+ 1+ 2
2q0 σ ρl
q0 γk
1< < +1 (R16)
ρlσ 2q0 σ
q0 2q0 σ
1− < (R17)
h0 D∞ kD∞

kD∞ q0
G5 (η) < √ 1− (R18)
πq0 σ h0 D ∞
exp(−x2 ) lk
= ,
x q0 cσ
G14 (η) > H14 (η) (R19)

q0
(1 − 2x2 ) = exp (x2 ).
ρlσ
Definition of functions (x > 0)

2q0 q0
F4 (x) = exp (−x ) − 1 x2 exp (−x2 )
2
γρcσ ρlσ
G4 (x) = x erf (x)
2q0 σ q0 cσ exp (−x2 )

F5 (x) = − 1 exp (−x2 )
γk lk x2
erf (x)
G5 (x) =
x
2q0 σ q0
f6 (x) = exp (−x ) exp (−x2 )
2
γk ρlσ
exp (−x2 )
G13 (x) =
erf (x)
√
2 πγ(1 − )
H13 (x) = x + b13 exp (x ) erf (x), con b13 =
2D∞ 1 − h0qD0 ∞

q0 2
G14 (x) = exp (−x ) − 1 x
ρlσ
H14 (x) = erf (x) exp (x2 )
Equations for ξ

σρcD∞ q0
G4 (x) = √ 1− x>0 (E4)
πq0 h0 D∞

kD∞ q0
G5 (x) = √ 1− x>0 (E5)
πq0 σ h0 D∞
√
πl
a13 G13 (x) = H13 (x), with a13 = x>0 (E13)
q0
cD∞ 1 − h0 D∞

2D∞ q0
a14 G14 (x) = H14 (x), with a14 =√ 1− x>0 (E14)
πγ(1 − ) h0 D∞
36
III. Phase-change problems for the fractional heat-diffusion equation
 Fractional derivatives
o Caputo fractional derivative
o Wright and Mainardi Functions
o Basic properties
 Two-phase fractional Stefan problems:

o Generalized Neumann solution with a temperature boundary condition
o Generalized Neumann solution with a heat flux boundary condition (In progress)
 One-phase fractional Stefan problem:

37
III.1 Fractional derivatives
Definition 1. Caputo fractional derivative (Caputo, Geophys. J. R. Astr. Soc. (1967)):
f ' (τ)
t
1
∫
α
D f (t) = dτ for 0 < α < 1
Γ(1 − α) 0 ( t − τ ) α
(D1)
= f ' (t) for α = 1
where
+∞
Γ(x) = ∫
0
t x −1 exp( − t) dt (Gamma function) . (D2)
Definition 2. Wright function (Wright, J. London Math. Soc. (1933)):
+∞
zn
W(z; α, β) = ∑ , z ∈ , α > −1, β ∈  . (D3)
n = 0 n! Γ (nα + β)
38
Definition 3. Mainardi function (Mainardi, Fract. Calc. Appl. Anal. (1999)):
+∞
(−z) n
M υ (z) = W( −z; −υ,1 − υ) = ∑ , z ∈ , υ < 1 (D4)
n = 0 n! Γ ( − nυ + 1 − υ)
which is a particular case of the Wright function.
Basic Properties:
∂W
(z; α, β) = W(z; α, α + β) . (D5)
∂z
1 x 1 x
W(− x; − ,1) = erfc   , 1 − W(− x; − ,1) = erf   . (D6)
2 2 2 2
Γ(1 + β) β−α
Dα ( t β ) = t (D7)
Γ(1 + β − α)
39
III.2 Generalized Neumann solutions
III.2.1. For Temperature boundary condition at the fixed face x = 0 (Roscani – T. (2014)).
III.2.2. For Heat Flux boundary condition at the fixed face x = 0 (Roscani – T., In
progress).
40
III.2.1) GENERALIZED NEUMANN SOLUTION (MELTING PROCESS) WITH A

TEMPERATURE BOUNDARY CONDITION
Problem ( FP1 ) Find the free boundary x = s(t ) , and the temperature T = T ( x , t ) such that the
following equations and conditions are satisfied (0 < α < 1) :
D Ts − s2Tsxx = 0, x > s(t ), t > 0 , (FT1)
D T − 2T xx = 0, 0 < x < s(t ), t > 0 , (FT2)
s(0) = 0 , (FT3)
Ts ( x ,0) = Ts ( +∞ , t ) = Ti < T f , x > 0, t > 0 , (FT4)
Ts (s(t ), t ) = T f , t > 0, (FT5)
Tl (s(t ), t ) = T f , t > 0, (FT6)
ksTsx ( s(t ), t ) − kT x ( s(t ), t ) =   D s(t ), t > 0 , (FT7)
T (0, t ) = T0 > T f , t > 0, (FT8)

ks k
where s2 = , =
2
 .
 cs  c
41
42
Theorem FT1 Let Ti < T f < T0 be. A solution of the problem ( FP1 ) is the generalized
Neumann explicit solution given by:
 x  
1 − W  −  ; − ,1 
 t 2 2  
T ( x , t ) = T0 − (T0 − T f )  
,  = s > 0, (FT9)
 
1 − W  − ; − ,1 
 2 
 x  
W  −  ; − ,1 
 t 2 2 
Ts ( x , t ) = Ti + (T f − Ti )  s
, (FT10)

W  − ; − ,1 
 2 

s(t ) =  st 2 , (FT11)
where the coefficient  =  > 0 is the solution of the following equation:
43
F (x) =
( )
Γ 1+ 2
x, x>0 (FT12)
Γ (1 − 2)


with
k (T0 − T f ) ks (T f − Ti )
F ( x ) = F1 ( x ) − F2 ( x ) , (FT13)
 s    s
2
and
M (x ) M (x )
F1 ( x ) = 2
, F2 ( x ) = 2
, (FT14)
 
1 − W  −x ; − ,1  W  −x ; − ,1 
 2   2 
where W and M α are the Wright and Mainardi functions respectively.

2
44
III.F2) GENERALIZED NEUMANN SOLUTION (MELTING PROCESS) WITH A

HEAT FLUX BOUNDARY CONDITION
Problem ( FP2 ) Find the free boundary x = s(t ) , and the temperature T = T ( x , t ) such that the
following equations and conditions are satisfied (0 < α < 1) :
D Ts − s2Tsxx = 0, x > s(t ), t > 0 , (FF1)
D T − 2T xx = 0, 0 < x < s(t ), t > 0 , (FF2)
s(0) = 0 , (FF3)
Ts ( x ,0) = Ts ( +∞ , t ) = Ti < T f , x > 0, t > 0 , (FF4)
Ts (s(t ), t ) = T f , t > 0, (FF5)
Tl (s(t ), t ) = T f , t > 0, (FF6)
ksTsx ( s(t ), t ) − kT x ( s(t ), t ) =   D s(t ), t > 0 , (FF7)

q
kT x (0, t ) = − 0 , t > 0, (FF8)
t 2
ks k
where s2 = , =
2
 .
 cs  c
45
46
NECESSARY CONDITION TO HAVE A FRACTIONAL TWO-PHASE PROBLEM

(MELTING PROCESS) WITH A HEAT FLUX BOUNDARY CONDITION
Problem ( FDEF2 ) Find the temperature T = T ( x , t ) such that the following equation and
conditions are satisfied (0 < α < 1) :
D T −  2Txx = 0, x > 0, t > 0 , (FDEF1)

T ( x ,0) = T ( +∞ , t ) = Ti < T f , x > 0, t > 0 , (FDEF2)
q0
kTx (0, t ) = −  , t > 0, (FDEF3)
t 2
k
where  2 = .
c
47
48
Theorem FDEF1 Let Ti < T f be.

a) The solution of the problem (FDEF1)- (FDEF3) is given by:
q0  Γ(1 −  2 )  x  
T ( x , t ) = Ti + W  −  ; − ,1  , x > 0, t > 0 . (FDEF4)
k  t 2 2 
b) The temperature at the boundary x = 0 is given by:
q0  Γ(1 −  2 )
T (0, t ) = Ti + = Const., ∀t > 0 . (FDEF5)
k
c) The necessary and sufficient condition in order to have an instantaneous fractional phase-
change is that the coefficient q 0 satisfies the inequality:
k (T f − Ti )
q0 > . (FDEF6)
 Γ(1 −  2 )
49
Theorem FF1 Let Ti < T f be.
a) If the coefficient q0 satisfies the inequality:
ks (T f − Ti )
q0 > , (FF9)
s Γ(1 −  2 )
then there exists an instantaneous phase-change (melting) process and the problem ( FP2 ) has
the generalized Neumann explicit solution given by:
q0  Γ(1 −  2 )   x     
T ( x , t ) = T f +  W  −  ; − ,1  − W  − F ; − ,1   , (FF10)
k   t 2 2   2 

 x  
W  −  ; − ,1 
 t 2 2 
Ts ( x , t ) = Ti + (T f − Ti )  s
, (FF11)

W  − F ; − ,1 
 2 

s(t ) =  F st 2 , (FF12)
50
where the coefficient  =  F > 0 is the solution of the following equation:
FF ( x ) =
(
Γ 1+ 2 ) x, x>0 (FF13)
Γ (1 −  2)
with
q0 Γ(1 −  2 ) ks (T f − Ti )
FF ( x ) = M ( x ) − F2 ( x ) . (FF14)
 s 2   s
2
b) If the coefficient q 0 satisfies the inequalities:

ks (T f − Ti )
0 < q0 ≤ , (FF15)

s Γ(1 − 2 )
then the problem ( FP2 ) is a fractional heat transfer problem for the initial solid phase whose
solution is given by:
q0 s Γ(1 −  2 )  x  
Ts ( x , t ) = Ti + W  −  ; − ,1  , x > 0, t > 0 . (FF16)
ks  2  
 st
2
51
Theorem FF2 Let Ti < T f be. If the coefficient q 0 verifies the inequality (FF9) then the
solution of the problem ( FP2 ) converges to the solution of the problem ( P2 ) when α → 1− and
we recover the solution and the inequality for the coefficient q 0 obtained for α = 1 in [T -
Quart. Appl. Math., 1981].
52
III.3 One-phase fractional Stefan problem
III.3.1 Determination of one unknown thermal coefficient through a one-phase fractional

Stefan problem with an over-specified condition on the fixed face x = 0 (T. Appl. Math.,
2015)
The method for determining unknown thermal coefficients through a one-phase fractional
Lamé-Clapeyron-Stefan with an over-specified condition on the fixed face x = 0 is given by:
Find the free boundary x = s(t ) , defined for t > 0 , and the temperature T = T ( x , t ) , defined
for x > 0 and t > 0 , such that the following equation and conditions are satisfied:
53

D T =  Txx ,2
0 < x < s(t ), t > 0 , (FC1)
s(0) = 0 , (FC2)
T ( x ,0) = T ( +∞ , t ) = Tm , x > 0, t > 0 , (FC3)
T (s(t ), t ) = Tm , t>0 , (FC4)
− kTx ( s(t ), t ) =   D s(t ), t > 0 , (FC5)
T (0, t ) = T0 > Tm , t > 0, (FC6)
q0
kTx ( 0, t ) = −  , t > 0, (FC7)
t 2
54
where:
 : density of mass,
k: thermal conductivity,
c: specific heat by unit of mass,
 : latent heat of fusion by unit of mass,
k
2 = > 0 : difussion coefficient,
c
T0 ( > Tm ) : temperature at the fixed face x = 0 .
q0 > 0 : coefficient that characterized the heat flux at the heat flux x = 0 , which must be
simultaneous experimentally determined with the temperature T0 .
The unknown thermal coefficient can be chosen among: k ,  , c and  .
55
Theorem 1 The solution to problem (FC1)-(FC7) with 0 < α < 1 and one unknown thermal
coefficient is given by:

s(t ) =   t 2 ,  > 0 , (FC8)
T0 − Tm   x  
T ( x , t ) = T0 − 1 − W  −  ; − ,1   , (FC9)
  t 2 2 
1 − W  − ; − ,1    
 2 
where the coefficient  > 0 and the unknown termal coefficient must satisfy the following
system of equations:
k (T0 − Tm )   
= 1 − W  − ; − ,1  , (FC10)
 2 
 q0 Γ  1 −  
 2
 
c (T0 − Tm ) Γ  1 −  1 − W  − ; − ,1 
 2  2 
= . (FC11)
 M ( )
 Γ  1 +  2
 2 
56
We have 4 cases:
Case 1: determination of { , k} ;
Case 2: determination of { , } ;
Case 3: determination of { , c } ;
Case 4: determination of { ,  } .
We show only the result for case # 3, and the Table 1 with the results of the 4 cases.
57
Theorem 2 (Case 3: Determination of thermal coefficient c)

If data satisfy the condition:
 
k   (T0 − Tm ) Γ  1 + 
 2
< 1, (FC12)

q02 Γ  1 − 
 2
then the solution to case 3 (problem (FC1)-(FC7) with 0 < α < 1 and the thermal unknown
coefficient c) is given by:
 
 Γ1 + 
 2     
c = c =  1 − W −
  ; − ,1   , (FC13)

(T0 − Tm ) Γ  1 −  M 2 ( )   2 
 2
where the coefficient  =  =  ( , k ,  ,  , q0 , T0 − Tm ) > 0 is the unique solution of the
equation:
 
1 − W  − x ; − ,1  k   (T0 − Tm ) Γ  1 + 
 2   2 1
= , x > 0. (FC14)
x  M ( x )
q02 Γ 3  1 −  2
 2 
58
Moreover, the temperature T ( x , t ) = T ( x , t ) and the free boundary s(t ) = s (t ) are given
by the following expressions ( 0 < α < 1):
T0 − Tm   x  
T ( x , t ) = T ( x , t ) = T0 − 1 − W  − ; − ,1   , (FC15)
   
  t 2 2 
1 − W  − ; − ,1     
 2 

s(t ) = s (t ) =   t 2 , (FC16)
where the coefficient  =  =  ( , k ,  ,  , q0 , T0 − Tm ) > 0 is the unique solution of the equation

(FC14) and the difusión coefficient  2 = 2 is given by:
 
k (T0 − Tm ) Γ  1 −  M ( )
k  2
 2 = 2 = = 2
. (24)
 c      
 Γ  1 +    1 − W −
  ; − ,1  
 2   2 
59
−
Theorem 3 If the parameter α → 1 then, under the hipothesis (FC12), the solution to case 3,
coincides with the one given in [T., Adv. Appl. Math. (1982)], that is:
1
s1 (t ) = 2 1 1 t 2 , 1 > 0 , (25)
T0 − Tm  x 
T1 ( x , t ) = T0 − erf  , (26)
erf ( 1 )  2 t 2 
1
 1 
 q02 k k (T0 − Tm )
c1 = erf 2 ( 1 ) , 1 = = , (27)
k  (T0 − Tm ) 2
 c 1 q0  erf ( 1 )
and the coefficient 1 > 0 is the unique solution of the following equation:
2 q02  erf ( x )
exp( x ) = , x > 0. (28)
k   (T0 − Tm ) x
In particular, the inequality (FC12) is transformed into the following one:
k   (T0 − Tm )
2
< 1. (29)
2 q0
60
Explicit formulae for the

Case Parameter  as the unique solution Restriction on
unknown thermal
# of the equation data
coefficient


q02 Γ 2  1 −  c (T0 − Tm ) Γ  1 − 
 
2
 2
1  2   F4 ( x ) = , x>0 -----------
k= 1 − W  − ; − ,1  
 c (T0 − Tm )  2    Γ  1 + 
2

 2

c (T0 − Tm ) Γ  1 −   ck (T0 − Tm )
 2 G3 ( x ) = , x>0
2 =   R2
  q0 Γ  1 − 
Γ  1 +  F4 ( ) 2
 2 
 

 Γ  1 +  1 − W  − x ; − ,1  k   (T0 − Tm ) Γ  1 + 
 2  2   2 1 R1
3 c= F ( ) = , x>0
 4
(T0 − Tm ) Γ  1 −  x 2 3  M ( x )
 2 q0 Γ  1 −  2
 2


q02 Γ 2  1 −  c (T0 − Tm ) Γ  1 − 
 
2
 2
4  2   F4 ( x ) = , x>0 -----------
= 1 − W  − ; − ,1  
kc (T0 − Tm )  2    Γ  1 + 
2

 2
Table 1. Summary of the determination of one unknown thermal coefficient through a one-
phase fractional Lamé-Clapeyron-Stefan with an over-specified condition on the fixed face (4
cases)
61
Restrictions on data:

k   (T0 − Tm ) Γ  1 + 
 2
R1: < 1;

q02 Γ  1 − 
 2
 ck (T0 − Tm )
R2: <1
 
q0 Γ  1 − 
 2
Real functions:
  
x 1 − W  −x ; − ,1  
 2 
F4 ( x ) =  , x>0
M ( x )
2
  
G3 ( x ) = 1 − W  −x ; − ,1  , x > 0
 2 
62
Case Explicit formulae for the Parameter  as the unique solution Restriction on
# unknown coefficient of the equation data
 q02 c (T0 − Tm )
k= erf 2 (  ) x exp( x 2 ) erf ( x ) = , x>0
1  c (T0 − Tm )
2
  -----------
c exp( −  2 ) erf ( x ) (T0 − Tm )  ck (T0 − Tm )  ck

2  = q0 = , x>0 <1
k  x q0  q0 
 q2 erf ( x ) k   (T0 − Tm )  k (T0 − Tm )

3 c= 0
erf 2 (  ) = exp( x 2 ), x > 0 <1
 k (T0 − Tm )
2
x q0 
2 2 q 20
 q02 c (T0 − Tm )
4 = erf 2 (  ) x exp( x 2 ) erf ( x ) = , x>0 -----------
kc (T0 − Tm )
2
 
Tabla 2. Summary of the determination of one unknown thermal coefficient through a one-
phase classical Lamé-Clapeyron-Stefan ( = 1 ) with an over-specified condition on the fixed
face (4 cases). These results can be obtained by taking the limit α → 1− in Table 1 (see T., Adv.
Appl. Math. (1982).
63
IV. Papers which are not reviewed
• Blengino – Reginto – T., Applied Math. Modelling (2015):
We propose a coupled water uptake-root growth model by using a moving boundary

problem for which the root length is prescribed as a function of time. The solution is obtained
by front-fixing and finite element methods.
• Gonzalez – Reginato - T., J. Biological Systems (2015):
We find a correct approximated solution using a polynomial of sixth degree for the free
boundary problem corresponding to the diffuion of oxygen in a spherical medium with
simultaneous absorption at a constant rate, and we show some mistakes in previously
published solutions.
64
Books on free boundary problems for the heat-diffusion equation

 V. Alexiades, A.D. Solomon, Mathematical modeling of melting and freezing processes, Hemisphere-Taylor &
Francis, Washington, 1996.
 J.R. Cannon, The one-dimensional heat equation, Addison-Wesley, Menlo Park, California, 1984.
 H.S. Carslaw, C.J. Jaeger, Conduction of heat in solids, Clarendon Press, Oxford, 1959.
 J. Crank, Free and moving boundary problem, Clarendon Press, Oxford, 1984.
 C.M. Elliott, J.R. Ockendon, Weak and variational methods for moving boundary problems, Research Notes in
Math. #59, Pitman, London, 1982.
 A. Fasano, Mathematical models of some diffusive processes with free boundary, MAT – Serie A, 11 (2005), 1-
128.
 S.C. Gupta, The classical Stefan problem. Basic concepts, modelling and analysis, Elsevier, Amsterdam, 2003.
 V.J. Lunardini, Heat transfer with freezing and thawing, Elsevier, London, 1991.
 L.I. Rubinstein, The Stefan problem, American Mathematical Society, Providence, 1971.
 D.A. Tarzia, A bibliography on moving-free boundary problems for heat diffusion equation. The Stefan problem,
MAT - Serie A, 2 (2000), 1-297.
 A.B. Tayler, Mathematical models in applied mechanics, Clarendon Press, Oxford, 1986.
 H. Weber, Die partiellen Differential-Gleinchungen der Mathematischen Physik, nach Riemann's Vorlesungen, t.
II, Braunwschweig, 1901.
65
Articles with explicit solutions on free boundary problems for the heat-diffusion
equation
 P. Boadbridge, Solution of a nonlinear absorption model of mixed saturated-unsaturated flow, Water Resources
Research, 26 (1990), 2435-2443.
 C. Briozzo, M.F. Natale, D.A. Tarzia, Explicit solutions for a two-phase unidimensional Lamé-Clapeyron-Stefan
problem with source terms in both phases, J. Math. Anal. Appl., 329 (2007), 145-162.
 A.C. Briozzo, D.A. Tarzia, Explicit solution of a free-boundary problem for a nonlinear absorption model of
mixed saturated-unsaturated flow, Adv. Water Resources, 21 (1998), 713-721.
 A.N. Ceretani, D.A. Tarzia, Similarity solutions for thawing processes with a convective boundary condition,
Rendiconti del’Istituto di Matematica dell’Università di Trieste, 46 (2014), 137-155.
 A.N. Ceretani, D.A. Tarzia, Determination of the one unknown thermal coefficient through a mushy zone model
with a convective overspecified boundary condition, Mathematical Problems in Engineering, Vol. 2015 Art ID
637852 (2015), 1-8.
 A.N. Ceretani, D.A. Tarzia, Simultaneous determination of the two unknown thermal coefficients through a
mushy zone model with an over-specified convective boundary condition, JP Journal of Heat and Mass Transfer,
13 No. 2 (2016), 277-301.
66
 R. Grzymkowsi, E. Hetmaniok, M. Pleszczynski, D. Slota, A certain analytical method used for solving the
Stefan problem, Thermal Science, 17 (2013), 635-642.
 G. Lamé, B.P. Clapeyron, Memoire sur la solidification par refroidissement d'un globe liquide, Annales Chimie
Physique, 47 (1831), 250-256.
 M.F. Natale, D.A. Tarzia, Explicit solutions to the two-phase Stefan problem for Storm's type materials, Journal
of Physics A: Mathematical and General, 33 (2000), 395-404.
 N.N. Salva, D.A. Tarzia, Explicit solution for a Stefan problem with variable latent heat and constant heat flux
boundary conditions, J. Math. Anal. Appl., 379 (2011), 240-244.
 A.D. Solomon, D.G. Wilson, V. Alexiades, Explicit solutions to change problems, Quart. Appl. Math., 41 (1983),
237-243.
 J. Stefan, Über einge probleme der theorie der Wärmeleitung, Zitzungberichte der Kaiserlichen Akademie der
Wissemschaften Mathematisch-Naturwissemschafthiche classe, 98 (1889), 473-484.
 D.A. Tarzia, An inequality for the coefficient  of the free boundary s(t ) = 2 t of the Neumann solution for the
two-phase Stefan problem, Quart. Appl. Math., 39 (1981), 491-497.
 D.A. Tarzia, An explicit solution for a two-phase unidimensional Stefan problem with a convective boundary
condition at the fixed face, MAT – Serie A, 8 (2004), 21-27.
 D.A. Tarzia, Explicit and approximated solutions for heat and mass transfer problems with a moving interface,
Chapter 20, in Advanced Topics in Mass Transfer, InTech Open Access Publisher, Rijeka, 2011, 439-484.
67
 D.A. Tarzia, Explicit solutions for the Solomon-Wilson-Alexiades’s mushy zone model with convective or heat
flux boundary conditions, J. Appl. Math., 2015 Art ID 375930 (2015), 1-9.
 D.A. Tarzia, Determination of the one unknown thermal coefficient through the one-phase fractional Lamé-
Clapeyron-Stefan problem, Appl. Math., 6 (2015), 2182-2191.
 D.A. Tarzia, elationship between Neumann solutions for two-phase Lamé-Clapeyron-Stefan problems with
convective and temperature boundary conditions, Thermal Science, In Press (2016). See arXiv:1406.0552
 V.R. Voller, F. Falcini, Two exact solutions of a Stefan problem with varying diffusivity, Int. J. Heat Mass
Transfer, 58 (2013), 80-85.
 V.R. Voller, J.B. Swenson, C. Paola, An analytical solution for a Stefan problem with variable latent heat, Int. J.
Heat Mass Transfer, 47 (2004), 5387-5390.
 S.M. Zubair, M.A. Chaudhry, Exact solutions of solid-liquid phase-change heat transfer when subjected to
convective boundary conditions, Heat Mass Transfer, 30 (1994), 77-81.
68
Books on fractional derivatives
 Kilbas, H. Srivastava, J. Trujillo, Theory and Applications of Fractional Differential Equations, Vol. 204 of
North-Holland Mathematics Studies, Elsevier, Amsterdam, 2006.
 F. Mainardi, Fractional calculus and waves in linear viscoelasticity, Imperial College Press, London, 2010.
 Podlubny, Fractional Differential Equations, Vol. 198 of Mathematics in Science and Engineering, Academic
Press, San Diego, 1999.
69
Articles on fractional derivatives

 M. Caputo, Linear model of dissipation whose Q is almost frequency independent - II, Geophys. J. R. Astr. Soc.,
13 (1967), 529-539.
 R. Gorenflo, Y. Luchko F. Mainardi, Analytical properties and applications of the Wright function, Fract. Calc.
Appl. Anal., 2 (1999), 383-414.
 Y. Luchko, Some uniqueness and existence results for the initial-boundary-value problems for the generalized
time-fractional diffusion equation, Computer and Mathematics with Applications, 59 (2010), 1766-1772.
 F. Mainardi, Y. Luchko, G. Pagnini, The fundamental solution of the space-time fractional diffusion equation,
Fract. Calc. Appl. Anal., 4 (2001), 153-192.
 F. Mainardi, A. Mura, G. Pagnini, The M-Wright function in time-fractional diffusion processes: a tutorial
survey, International Journal of Differential Equations, Vol. 2010, Article ID 104505, 1-29.
 E.M. Wright, On the coefficients of power series having exponential singularities, J. London Math. Soc., 8
(1933), 71-79.
 E.M. Wright, The asymptotic expansion of the generalized Bessel function, J. London Math. Soc., 10 (1935), 287-
293.
 E.M. Wright, The generalized Bessel function of order greater than one, Quart. J. Math., 11 (1940), 36-48.
70
Articles with fractional derivatives on the Lamé-Clapeyron-Stefan problem
 C. Atkinson, Moving boundary problems for time fractional and composition dependent diffusion, Fract. Calc.
Appl. Anal., 15 (2012), 207-221.
 F. Falcini, R. Garra, V.R. Voller, Fractional Stefan problems exhibing lumped and distributed latent-heat memory
effects, Physical Review E, 87 (2013), 042401, 1-6.
 L. Jinyi, X. Mingyu, Some exact solutions to Stefan problems with fractional differential equations, J. Math.
Anal. Appl., 351 (2009), 536-542.
 L.P. Kholpanov, Z.E. Zaklev, V.A. Fedotov, Neumann-Lamé-Clapeyron-Stefan Problem and its solution using
Fractional Differential-Integral Calculus, Theoretical Fundations of Chemical Engineering, 37 (2003), 113-121.
 S. Roscani, E.A. Santillan Marcus, Two equivalent Stefan's problems for the time-fractional diffusion equation,
Fract. Calc. Appl. Anal., 16 (2013), 802-815.
 S. Roscani, E.A. Santillan Marcus, A new equivalence of Stefan's problems for the time-fractional diffusion
equation, Fract. Calc. Appl. Anal., 17 (2014), 371-381.
 S. Roscani, D.A. Tarzia, A generalized Neumann solution for the two-phase fractional Lamé-Clapeyron-Stefan
problem, Adv. Math. Sci. Appl. 24 (2014), 237-249.
Clapeyron-Stefan problem, App. Math.s, 6 (2015), 2182-2191.
 V.R. Voller, An exact solution of a limit case Stefan problem governed by a fractional diffusion equation, Int. J.
Heat and Mass Transfer, 53 (2010), 5622-5625.
 V.R. Voller, Fractional Stefan problems, Int. J. Heat and Mass Transfer, 74 (2014), 269-277.
71
OUR RECENT PUBLICATIONS ON JOURNALS

 J. L. Blengino Albrieu, J.C. Reginato, D.A. Tarzia, Modeling water uptake by a root system growing in a fixed
soil volume, Applied Mathematical Modelling, 39 (2015), 3434-3447.
 A.C. Briozzo, D.A. Tarzia, Convergence of the solution of the one-phase Stefan problem with respect two
parameters, MAT – Serie A, 20 (2015), 31-38.
 A.N. Ceretani, D.A. Tarzia, Similarity solutions for thawing processes with a convective boundary condition,
Rendiconti del’Istituto di Matematica dell’Università di Trieste, 46 (2014), 137-155.
 A.N. Ceretani, D.A. Tarzia, L.T. Villa, Explicit solutions for a non-classical heat conduction problem for a semi-
infinite strip with a non-uniform heat source, Boundary Value Problems, 2015 No. 156 (2015), 1-26.
 A.N. Ceretani, D.A. Tarzia, Determination of the one unknown thermal coefficient through a mushy zone model
with a convective overspecified boundary condition, Mathematical Problems in Engineering, Vol. 2015 Art ID
637852 (2015), 1-8.
 A.N. Ceretani, D.A. Tarzia, Simultaneous determination of the two unknown thermal coefficients through a
mushy zone model with an over-specified convective boundary condition, JP Journal of Heat and Mass Transfer,
13 No. 2 (2016), 277-301.
 A.M. Gonzalez, J.C. Reginato, D.A. Tarzia, A free boundary problem for oxygen diffusion in a sphere, Journal of
Biological Systems, 23 Supp 01 (2015), S67-S76.
 S.D. Roscani, D.A. Tarzia, A generalized Neumann solution for the two-phase fractional Lamé-Clapeyron-Stefan
problem, Advances in Mathematical Sciences and Applications, 24 No. 2 (2014), 237-249.
 D.A. Tarzia, Explicit solutions for the Solomon-Wilson-Alexiades’s mushy zone model with convective or heat
flux boundary conditions, Journal of Applied Mathematics, 2015 Art ID 375930 (2015), 1-9.
Clapeyron-Stefan problem, Applied Mathematics, 6 (2015), 2182-2191.
 D.A. Tarzia, Relationship between Neumann solutions for two-phase Lamé-Clapeyron-Stefan problems with
convective and temperature boundary conditions, Thermal Science, In Press (2016). See arXiv:1406.0552
72
WORK IN PROGRESS
1) J. BOLLATI – A.N. CERETANI – S.D. ROSCANI - D.A. TARZIA, “The two-phase fractional Lamé-Clapeyron-
Stefan problem with a convective boundary condition”.
2) J. BOLLATI – D.A. TARZIA, “Explicit solution for the one-phase Stefan problem with latent heat depending on
the position and a convective boundary condition by using Kummer functions”.
3) A.N. CERETANI – N. SALVA - D.A. TARZIA, “The one phase Stefan problem with a temperature-depended
thermal conductivity and a convective boundary condition”.
4) A.N. CERETANI – D.A. TARZIA, "Simultaneous determination of the two unknown thermal coefficients
through the one-phase fractional Lamé-Clapeyron-Stefan problem”.
5) A.N. CERETANI – D.A. TARZIA, "Similarity solution for a two-phase Stefan problem with a convective
boundary condition and a mushy model”.
6) J.C. REGINATO – J. L. BLENGINO ALBRIEU – D.A. TARZIA, "Analysis and use of cumulative nutrient
uptake formulas in plant nutrition and the temporal weight averaged influx”.
7) S.D. ROSCANI - D.A. TARZIA, “The two-phase fractional Lamé-Clapeyron-Stefan problem with a heat flux
boundary condition”.
8) S.D. ROSCANI - D.A. TARZIA, “Integral relation between temperature and the free boundary in the one-phase
fractional Lamé-Clapeyron-Stefan problem with a temperature boundary condition”.
73
Aknowledgements:
The present work has been partially sponsored by the Project PIP No 0534 from
CONICET – Univ. Austral, Rosario, Argentina, and Grant AFOSR-SOARD FA9550-14-1-
0122.
Collaborators:
• Dr. Andrea Ceretani (CONICET & Univ. Austral), Rosario, Argentina. Postdoc Fellowship;
• Dr. Juan C. Reginato (Univ. Nac. Río Cuarto), Río Cuarto, Argentina;
• Dr. Sabrina Roscani (CONICET & Univ. Austral), Rosario, Argentina. Postdoc Fellowship;
• Julieta Bollati (CONICET & Univ. Austral), Rosario, Argentina. Doc Fellowship, PhD. Student (2nd
year).
THANK YOU VERY MUCH
FOR YOUR ATTENTION

Explicit Solutions To Phase-Change Problems and Applications

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Explicit Solutions To Phase-Change Problems and Applications

Uploaded by

Copyright:

Available Formats

2016 Computational Mathematics Annual Program Review

Arlington, VA (USA), August 8-11, 2016

EXPLICIT SOLUTIONS TO PHASE-CHANGE

Domingo Alberto Tarzia

Grant AFOSR – SOARD FA 9550-14-1-0122

CONICET - Depto. Matemática, FCE, Univ. Austral,

Mathematics Subject Classification 2010: 26A33, 35C05, 35R11, 35R35, 80A22.

• Simultaneous determination of two unknown thermal coefficients through a

i) Continuous casting of steel (e.g., ingots, slabs);

and some related free boundary problems as:

xiii) Oxygen diffusion-consumption in a tissue (equivalent to the phase-change problem);

Obtaining explicit solution we have:

II. Classical one-phase Stefan problem with a simple mushy region

III. Phase-change problems for the fractional heat-diffusion equation

 Two-phases fractional Stefan problems:

 One-phase fractional Stefan problem:

IV. Papers which are not reviewed

I. Classical two-phase Stefan problems

I.1) CLASSICAL NEUMANN SOLUTION (MELTING PROCESS) WITH A

I.2) CLASSICAL NEUMANN SOLUTION (MELTING PROCESS) WITH A HEAT

Theorem F1 [T., Quart. Appl. Math. (1981)] Let Ti < T f be.

a) If the coefficient q0 satisfies the inequality:

b) If the coefficient q0 satisfies the inequalities:

I.3) CLASSICAL NEUMANN SOLUTION (SOLIDIFICATION PROCESS) WITH A

b) If the coefficient h0 satisfies the following inequalities:

II. Classical one-phase Stefan problem with a simple mushy region

 One-phase with a convective boundary condition (T., J. Appl. Math. (2015))

o Determination of one unknown thermal coefficient (Ceretani-T., Math. Problems

o Simultaneous determination of two unknown thermal coefficients (Ceretani-T., JP

Solidification process with a convective boundary condition:

 cTt − kTxx = 0 , 0 < x < s ( t ) , t > 0 ( = k /  c ) (SC1)

kTx (s(t ), t ) =  [ s(t ) + (1 −  )r(t )], t > 0 (SC3)

Tx (s(t ), t )(r (t ) − s(t )) =  > 0, t > 0 (with  > 0) (SC4)

s(0) = r (0) = 0 (SC5)

then the solution to the problem (SC1)-(SC6) is given by:

s(t ) = 2  t , t > 0 , (SC9)

where the coefficient  is given as the unique solution of the equation:

Solidification process with a heat flux condition:

 cTt − kTxx = 0 , 0 < x < s ( t ) , t > 0 ( = k /  c ) (SF1)

kTx (s(t ), t ) =  [ s(t ) + (1 −  )r(t )], t > 0 (SF3)

Tx (s(t ), t )(r (t ) − s(t )) =  > 0, t > 0 (with  > 0) (SF4)

s(0) = r (0) = 0 (SF5)

then the solution to the problem (SF1)-(SF6) is given by:

s(t ) = 2  t , t > 0 , (SF9)

where the coefficient  is given as the unique solution of the equation:

III. Determination of unknown thermal coefficients through a one-phase Stefan problem

kTx (s(t ), t ) =  [ s(t ) + (1 −  )r(t )], t > 0 (1FC3)

Tx (s(t ), t )(r (t ) − s(t )) =  , t > 0 (1FC4)

The problem can be divided in 6 different cases:

Case # One Unknown Coefficient

1 Latent heat of fusion by unit of mass: 

2 First parameter of the mushy region: 

3 Second parameter of the mushy region: 

6 Specific heat by unit of mass: c

Case Thermal coefficient Coefficient ξ that characterizes Restrictions

Definitions of functions (x > 0)

f2 (x) = x exp (x2 )

III.2 Simultaneous Determination of two thermal coefficients

where  > 0 is a positive parameter obtained experimentally.

 cTt − kTxx = 0 , 0 < x < s ( t ) , t > 0 ( = k /  c ) (2FC2)

kTx (s(t ), t ) =  [ s(t ) + (1 −  )r(t )], t > 0 (2FC4)

Tx (s(t ), t )(r (t ) − s(t )) =  , t > 0 (2FC5)

s(0) = r (0) = 0 (2FC6)

1 0<<1 – (R0), (R1)

4 = 1 − F4 (ξ) (E4) (R2), (R3), (R4),

5 = 1 − F5 (ξ) (E5) (R12)

6 = 1 − F6 (ξ) (E5) (R13), (R14), (R15)