4.15. Shape Memory Alloy (SMA)

A shape memory alloy (SMA) is a metallic alloy that can undergo permanent deformation at low temperature and then recover its original shape through a phase transformation activated by increasing temperature. Upon loading and unloading cycles, an SMA can exhibit the following types of behavior:

  • Superelasticity (also known as pseudoelasticity) – Large deformation without residual strains.

    Superelasticity and the shape memory effect occur due to the material microstructure in which two different crystallographic structures exist: austenite and martensite. Austenite is the crystallographically more-ordered phase, and martensite is the crystallographically less-ordered phase. The key characteristic of an SMA is the occurrence of a martensitic phase transformation.
  • Shape memory effect – The material recovers its original shape through thermal cycles.

    Typically, austenite is stable at high temperatures and low stress, while martensite is stable at low temperatures and high stress. The reversible martensitic phase transformation results in unique effects: the pseudoelasticity (PE) and the shape memory effect (SME).
  • Plasticity – A permanent strain that remains in the unloaded state.

    Similar to plasticity in other metals, plastic deformation in SMA materials occurs in either the austinite or martensite phase when the stress reaches the yield stress. The SMA phase transformation and plasticity are coupled through the martensite volume fraction.

The following topics about the SMA material model are available:

Also see Material Model Support for Elements for SMA. Element support varies according to the SMA material model option.

4.15.1. SMA Material Model Options

SMA material model options are available for:

  • Simulating superelastic behavior – Based on Auricchio et al.,[1] where the material undergoes large-deformation without showing permanent deformation under isothermal conditions.

  • Simulating the shape memory effect behavior of shape memory alloys – Based on the 3D thermomechanical model for stress-induced solid phase transformations of Souza et al.,[2] Auricchio et al.,[3] and Auricchio and Petrini.[4]

  • Combining plasticity with superelasticity and shape memory effect – A modified form of the Hartl and Lagoudas [7] model.

4.15.1.1. SMA Model Option: Superelasticity

4.15.1.1.1. Constitutive Model for Superelasticity

From a macroscopic perspective, the phase-transformation mechanisms involved in superelastic behavior are:

  1. Austenite to martensite (A->S)

  2. Martensite to austenite (S->A)

  3. Martensite reorientation (S->S)

Figure 4.57: Typical Superelasticity Behavior

Typical Superelasticity Behavior

Two of the phase transformations are considered here: A->S and S->A. The material is composed of two phases, the austenite (A) and the martensite (S). Two internal variables, the martensite fraction (ξS) and the austenite fraction (ξA), are introduced. One of them is a dependent variable, and they are assumed to satisfy the relation expressed as:

The independent internal variable chosen here is ξS.

The material behavior is assumed to be isotropic. The pressure dependency of the phase transformation is modeled by introducing the Drucker-Prager loading function, as follows:

where α is the material parameter, σ is the stress, and 1 is the identity tensor.

The evolution of the martensite fraction, ξS, is then defined as follows:

where:

where are the material parameters shown in the following figure:

Figure 4.58: Idealized Stress-Strain Diagram of Superelastic Behavior

Idealized Stress-Strain Diagram of Superelastic Behavior

where are the material parameters shown in Figure 4.58: Idealized Stress-Strain Diagram of Superelastic Behavior.

The material parameter α characterizes the material response in tension and compression. If tensile and compressive behaviors are the same, then α = 0. For a uniaxial tension-compression test, α can be related to the initial value of austenite to martensite phase transformation in tension and compression (, respectively) as:

The stress-strain relation is:

where D is the elastic stiffness tensor, is the transformation strain tensor, and is the material parameter shown in Figure 4.58: Idealized Stress-Strain Diagram of Superelastic Behavior.

To model the difference between the austenite and martensite elastic properties, the elastic modulus (used in the material constitutive model) is a function of the martensite fraction , the elastic modulus of pure austenite (the initial input elastic modulus), and the elastic modulus of pure martensite :

4.15.1.1.2. Material Parameters for the Superelastic SMA Material Model

To model the superelastic behavior of shape memory alloys, initialize the data table via the TB,SMA command's SUPE option.

Define the elastic behavior in the austenite state (MP).

The superelastic SMA option is described by six constants that define the stress-strain behavior in loading and unloading for the uniaxial stress-state.

For each data set, define the temperature (TBTEMP), then define constants C1 through C7 (TBDATA). You can define up to 99 sets of temperature-dependent constants in the same way.

Table 4.30: Superelastic Option Constants

Constant Meaning Property
C1 Starting stress value for the forward phase transformation
C2 Final stress value for the forward phase transformation
C3 Starting stress value for the reverse phase transformation
C4 Final stress value for the reverse phase transformation
C5 Maximum residual strain
C6 α Parameter measuring the difference between material responses in tension and compression
C7 Elastic modulus of the full martensite phase

If 0 or undefined, the martensite and austenite phases share the same elastic modulus


Example 4.38: Defining Elastic Properties of the Austenite Phase

MP,EX,1,60000.0
MP,NUXY,1,0.36
Define SMA material properties
TB,SMA,1,,,SUPE
TBDATA,1, 520, 600, 300, 200, 0.07, 0.0

4.15.1.2. SMA Mode Option: Shape Memory Effect

4.15.1.2.1. The Constitutive Model for Shape Memory Effect

The shape memory effect was based on a 3D thermomechanical model for stress-induced solid phase transformations that was presented in [2] [3][4]. Within the framework of classical irreversible thermodynamics, the model is able to reproduce all of the primary features relative to shape memory materials in a 3D stress state. The free energy potential is set to:

where:

D = material elastic stiffness tensor
= total strain
= total transformation strain
= deviatoric transformation strain
τM(T) = a positive and monotonically increasing function of the temperature as 〈β(T - T0)〉+ in which 〈∙〉+ is the positive part of the argument (also known as Maxwell stress).
β = material parameter
T = temperature
T0 = phase-balance temperature
h = material parameter related to the hardening of the material during the phase transformation
= indicator function introduced to satisfy the constraint on the transformation norm [1] in which

from which we have

where X tr is defined as the transformation stress.

Stresses, strains, and the transformation strains are then related as follows:

Splitting the stress into deviatoric and volumetric components, we have

where S is the deviatoric stress and p is the volumetric stress (also called hydrostatic pressure)

The transformation stress is given as follows:

where γ is defined by

where is a maximum value of .

Numerous experimental tests show an asymmetric behavior of SMA in tension and compression, and suggest describing SMA as an isotropic material with a Prager-Lode-type limit surface. Accordingly, the following yield function is assumed:

where X tr is the transformation stress, J2 and J3 are the second and third invariants of transformation stress, m is a material parameter related to Lode dependency, and R is the elastic domain radius.

J2 and J3 are defined as follows:

The evolution of transformation strain is defined as:

where ξ is an internal variable and is called as transformation strain multiplier. ξ and F(X tr) must satisfy the classical Kuhn-Tucker conditions, as follows:

which also reduces the problem to a constrained optimization problem.

The elastic properties of austenite and martensite phase differ. During the transformation phase, the elastic stiffness tensor of material varies with the deformation. The elastic stiffness tensor is therefore assumed to be a function of the transformation strain , defined as:

where D A is the elastic stiffness tensor of austenite phase, and D S is the elastic stiffness tensor of martensite phase. The Poisson’s ratio of the austenite phase is assumed to be the same as the martensite phase. When the material is in its austenite phase, D = D A, and when the material undergoes full transformation (martensite phase), D = D S.

The following figure illustrates a number of the mechanical model features:

Figure 4.59: Admissible Paths for Elastic Behavior and Phase Transformations

Admissible Paths for Elastic Behavior and Phase Transformations

The austenite phase is associated with the horizontal region abcd. Mixtures of phases are related to the surface cdef. The martensite phase is represented by the horizontal region efgh. Point c corresponds to the nucleation of the martensite phase. Phase transformations take place only along line cf, where . Saturated phase transformations are represented by paths on line fg. The horizontal region efgh contains elastic processes except, of course, those on line fg.

A backward Euler integration scheme is used to solve the stress update and the consistent tangent stiffness matrix required by the finite element solution for obtaining a robust nonlinear solution. Because the material tangent stiffness matrix is generally unsymmetric, use the unsymmetric Newton-Raphson option (NROPT,UNSYM) to avoid convergence problems.

Assuming that A and S share the same elastic modulus (DS = DA) in a uniaxial loading under the constant temperature T, these are the four critical stress points (shown in Figure 4.60: Stress Points at the Start/End of the Phase Transfers) at the start and end of the phase transformations:

 

Figure 4.60: Stress Points at the Start/End of the Phase Transfers

Stress Points at the Start/End of the Phase Transfers

4.15.1.2.2. Material Parameters for the Shape Memory Effect Option

To model the shape memory effect behavior of shape memory alloys, initialize the data table via the TB,SMA command’s MEFF option.

Define the elastic behavior in the austenite state (MP).

The shape memory effect option is described by seven constants that define the stress-strain behavior of material in loading and unloading cycles for the uniaxial stress-state and thermal loading.

For each data set, define the temperature (TBTEMP), then define constants C1 through C7 (TBDATA). You can define up to 99 sets of temperature-dependent constants in this manner.

Table 4.31: Shape Memory Effect Option Constants

Constant Meaning Property

C1

h

Hardening parameter

C2

To

Phase-balance temperature

C3

R

Elastic limit

C4

β

Temperature scaling parameter

C5

Maximum value of , where is the deviatoric transformation strain

C6

Em

Martensite modulus

C7

m

Lode dependency parameter


Example 4.39: Defining Shape Memory Effect Properties of the Austenite Phase

MP,EX,1,60000.0
MP,NUXY,1,0.36
Define SMA material properties
TB,SMA,1,,,MEFF
TBDATA,1,1000, 223, 50, 2.1, 0.04, 45000
TBDATA,7,0.05

4.15.1.3. SMA Model Option: Shape Memory Effect with Plasticity

This SMA model includes superelasticity, the shape memory effect, and plastic yielding. Following a small-strain formulation, the total strain is additively decomposed through:

in terms of the elastic strain , thermal strain , transformation strain , and plastic strain . The transformation strain is decomposed into a magnitude and direction through:

where:

= material parameter characterizing the maximum transformation strain
= volume fraction of martensite constrained to the range of [0,1], representing the formation and depletion of all martensite variants
= direction of the transformation strain constrained to have a unit magnitude (that is, )

The foundation of the model is the Helmholtz free-energy function, assumed to take the form:

The elastic part contains:

= phase-dependent bulk modulus
= total volumetric strain
= phase-dependent shear modulus
= total deviatoric strain
= coefficient of thermal expansion
= temperature (absolute). (The program adds the offset temperature [TOFFST] to all temperatures and corresponding coefficients.)
= reference temperature for thermal expansion

The thermal energy contains:

= phase-balance temperature (at which the specific free energies of the two phases are the same at stress-free conditions) (absolute)
is an internal variable providing a lower bound for
= entropy difference during transformation

Note:   during forward transformation, and during reverse transformation.


The interaction part contains:

= hardening energy
= transformation hardening parameter

The constraint energy is defined using:

= Lagrange multipliers (which enforce the constraints and , respectively)

The plastic energy contains:

= linear kinematic hardening modulus (phase-independent)

Both the bulk and shear moduli are dependent on the phase through a Reuss homogenization approach using:

with representing the value in austenite and the value in martensite.

The definitions of can be obtained from the material parameters through the following:

Using the Helmholtz free-energy function, the stress is:

Considering a constant , the dissipation function for transformation is assumed to take on the form of:

where represents the radius of the elastic domain of transformation. The parameter controls the width of the hysteresis of the outer loop, while changes the hysteresis for the internal loops. The factor modifies the radius of the elastic domain based on the amount of phase available for transformation. The available phase is defined based on the memory points as for reverse transformation and for forward transformation.

The transformation hardening energy is defined through its derivative with respect to :

additionally accounts for the smooth development of internal (or minor) loops in the case of incomplete transformations using:

with the hardening function for the outer (or major) loop , the hardening function for the internal loop and a factor for adjusting the internal loop feature. The outer loop hardening is described using a polynomial function:

Here, the factors and describe the offset of the transformation stress from that of a linear response near the beginning and end of transformation, and is a linear hardening factor. and affect the curvature of the hardening process near the beginning and end of transformation.

The internal loop hardening function is represented based on a scaling of the outer loop hardening based on the memory points and . During forward transformation the internal loop factor and the internal loop transformation hardening are obtained as

Similarly, during reverse transformation the relevant variables are defined by

The internal loop factor depends on a material parameter for turning the internal loop behavior on or off.

Figure 4.61: Effects of Effects of and on the Behavior of the Internal Loops and Effects of and on the Behavior of the Internal Loops on the Behavior of the Internal Loops

Effects of and on the Behavior of the Internal Loops

The thermodynamic force associated with plasticity is defined as:

The limit function for the plasticity is considered to be phase and stress-state dependent through the expressions:

Here, is a nonlinear isotropic hardening function decomposed into an austenite part and a martensite part. is a function to introduce tension-compression asymmetry through a dependence on the transformation strain direction . The material parameter defines the factor of increase in the isotropic hardening and yield point for the plasticity in compression compared to tension. For tension-compression symmetric behavior, is set to 1. The nonlinear isotropic hardening functions are dependent on the phase using:

where is the initial yield point of austenite and is the initial yield point of martensite. The isotropic hardening becomes saturated to the yield point for austenite and for martensite. The isotropic linear hardening following the exponential hardening is defined through and for austenite and martensite, respectively. The parameter determines how rapidly the saturation from to and to takes place. is the accumulated equivalent plastic strain.

Figure 4.62: Plastic Response of Austenite

Plastic Response of Austenite

The function is the basis for the tension-compression asymmetry and is defined as:

where the internal parameters , , and define the tension-compression asymmetric surface and . The internal parameters are chosen such that the function can achieve maximum asymmetry while remaining convex. The maximum transformation strain can be made tension-compression asymmetric using:

in terms of the maximum transformation strain in tension and the maximum transformation strain in compression . Similarly, the Young’s modulus of martensite can be made tension-compression asymmetric using:

in terms of the Young's modulus of martensite in tension and compression .

Figure 4.63: Stress-strain curve displaying tension-compression behavior of the maximum transformation strain and the Young’s modulus of martensite

Stress-strain curve displaying tension-compression behavior of the maximum transformation strain and the Young’s modulus of martensite

If values for the corresponding parameters are not defined, then default values are used so that the tension and compression behavior of the material are the same.

Note that the tension-compression asymmetry feature is optional. If this feature is not enabled, the model is symmetric in tension and compression. Tension-compression asymmetry is activated for the maximum transformation strain and the Young's modulus of martensite by defining and , respectively. In that case, the input parameters and are set to and to , respectively. The internal loop behavior is also optional.

The material tangent-stiffness matrix is unsymmetric for this material model, so it can be beneficial to use the unsymmetric Newton-Raphson solution option (NROPT).

4.15.1.3.1. Material Parameters for the Shape Memory Effect and Plastic Deformation

The austenite phase elastic behavior is defined with the material elastic constants (TB or MP). Initialize the data table (TB,ELAS) and define the Young's modulus of austenite ( ) and the Poisson's ratio ( ). Alternatively, use MP,EX to define the Young's modulus of austenite and MP,PRXY/NUXY to define the Poisson's ratio. The austenite phase elastic behavior is assumed to be isotropic.

To define the constants, initialize the data table (TB,SMA) with options (TBOPT) METE, METL, METH, MEPD, and METC. The function of each option follows:

METE (required) -- Defines the martensite phase elastic response and thermal expansion.
METL (required) -- Specifies the limits of transformation in strain-stress-temperature space and thus defines the phase diagram of a given SMA material.
METH (required) -- Considers the smooth behavior of transformation by applying a hardening function for the evolution of the martensite volume fraction .
MEPD (required) -- Defines the plastic phase-dependent yielding response. The plasticity exhibits linear kinematic and nonlinear isotropic hardening.
METC (optional) -- Defines the tension-compression asymmetry response and hysteresis response. If not defined, the program determines default parameter values.

For each data set, define temperature- and field-dependence (TBFIELD), if any, then define the material parameters (TBDATA) as shown in the following tables:

Table 4.32: TBDATA Constants for TB,SMA,,,,TBOPT = METE

ConstantMeaningPropertyRange
C1 Young’s modulus of martensite
C2 Thermal expansion coefficient

Table 4.33: TBDATA Constants for TB,SMA,,,,TBOPT = METL

ConstantMeaningPropertyRange
C1 Maximum transformation strain
C2 Phase-transformation limit
C3 Phase-balance temperature
C4 Entropy difference for forward transformation
C5 Entropy difference for reverse transformation

Table 4.34: TBDATA Constants for TB,SMA,,,,TBOPT = METH

ConstantMeaningPropertyRange
C1 Transformation hardening modulus
C2 Transformation hardening factor
C3 Transformation hardening factor
C4 Transformation hardening factor
C5 Transformation hardening exponent
C6 Transformation hardening exponent

Table 4.35: TBDATA Constants for TB,SMA,,,,TBOPT = MEPD

ConstantMeaningPropertyRange
C1 Isotropic hardening modulus of austenite
C2 Isotropic hardening modulus of martensite
C3 Isotropic plastic yield point of austenite
C4 Isotropic plastic yield point of martensite
C5Isotropic plastic yield saturation point of austensite
C6 Isotropic plastic yield saturation point of martensite
C7 Isotropic plastic yield saturation rate parameter
C8Kinematic hardening modulus

Table 4.36: TBDATA Constants for TB,SMA,,,,TBOPT = METC

ConstantMeaningPropertyRangeDefault
C1Young’s modulus of martensite in compression
C2Maximum transformation strain in compression
C3Compression plasticity factor (the ratio of the compression-yield point to the tension-yield point)

1

C4Hardening internal loop factor

0

C5Hysteresis internal loop factor

0


Example 4.40: Defining the Shape Memory Effect Properties of the Austenite Phase

!================================================!
! Material parameters for SMA plasticity model
!================================================!
! Thermo-elasticity
C1 = 37700      !EA
C2 = 31800      !ES
C3 = 0.33       !nu
C4 = 0.0        !alpha
! Transformation limits
C5 = 0.041      !eL
C6 = 259        !YS
C7 = 261        !theta0
C8 = 6.4        !deletaf
C9 = 9.4        !deletar
! Transformation hardening
C10= -9000      !mut
C11= -70        !a1
C12= 300        !a2
C13= 9500       !a3
C14= 30         !n1
C15= 30         !n2
! Plasticity
C16= 1000       !hpA
C17= 1000       !hpS
C18= 1000       !YpA
C19= 1000       !YpS
C20= 1200       !YmA
C21= 1200       !YmS
C22= 200        !CH
C23= 14000      !mup

TB,ELAS,1
TBDATA,1,C1,C3
TB,SMA,mat1,,,METE
TBDATA,1,C2,C4
TB,SMA,mat1,,,METL
TBDATA,1,C5,C6,C7,C8,C9
TB,SMA,mat1,,,METH
TBDATA,1,C10,C11,C12,C13,C14,C15
TB,SMA,mat1,,,MEPD
tbdata,1,C16,C17,C18,C19,C20,C21
tbdata,7,C22,C23

4.15.1.3.1.1. Specifying Temperature Units

The default temperature unit is Kelvin. If using another temperature unit, the corresponding coefficients for the constants need to be converted. This table shows how to convert a given coefficient value from one temperature-unit system to another:

Table 4.37: Example: Converting Coefficient Values

ConstantKelvin (Default)CelsiusFarenheit

4.15.2. Result Output of Solution Variables

For postprocessing, solution output is as follows:

  • Stresses are output as S.

  • Elastic strains are output as EPEL.

  • For the superelasticity (SUPE) and shape memory effect (MEFF) SMA model options:

    • Transformation strains are output as EPPL.

    • The value of is available as part of nonlinear solution record NL and can be processed as component EPEQ of NL.

  • For the shape memory effect with plastic deformation (METE+METL+METH+MEPD) SMA model option:

    • EPPL represents the plastic strain.

    • Transformation strain is output as EPCR.

    • EPEQ of NL represents the accumulative equivalent plastic strain.

  • Elastic strain energy density is available as part of the strain energy density record SEND (ELASTIC).

4.15.3. References

  1. Auricchio, F. (2001). A robust integration-algorithm for a finite-strain shape-memory-alloy. International Journal of Plasticity. 17, 971-990.

  2. Souza, A. C., Mamiya, E. N., & Zouain, N. (1998). Three-dimensional model for solids undergoing stress-induced phase transformations. European Journal of Mechanics-A/Solids. 17, 789-806.

  3. Auricchio, F., Taylor, R. L., & Lubliner, J. (1997). Shape-memory alloys: Macromodeling and numerical simulations of the superelastic behavior. Computational Methods in Applied Mechanical Engineering. 146(1), 281-312.

  4. Auricchio, F. & Petrini, L. (2005). Improvements and algorithmical considerations on a recent three-dimensional model describing stress-induced solid phase transformations. International Journal for Numerical Methods in Engineering. 55, 1255-1284.

  5. Auricchio, F., Fugazza, D., DesRoches, R. (2006). Numerical and experimental evaluation of the damping properties of shape-memory alloys. Journal of Engineering Materials and Technology. 128(3), 312-319.

  6. Hartl, D. J., & Lagoudas, D. C. (2009). Constitutive modeling and structural analysis considering simultaneous phase transformation and plastic yield in shape memory alloys. Smart Materials and Structures. 18(10), 1-17.

  7. Woodworth, L.A., Wang, X., Lin, G., & Kaliske, M. (2022). A multi-featured shape memory alloy constitutive model incorporating tension-compression asymmetric interpolation. Mechanics of Materials. 172.

For an example analysis, see Shape Memory Alloy (SMA) with Thermal Effect in the Technology Showcase: Example Problems.