Softening Laws

Introduction

Softening laws are used in softening materials to determine crack tractions as a function of effective crack opening displacements. Each law defines a normalized function that decays from 1 to 0 as damage develops up to the point of failure. The area under the law (times an associated stress) gives the fracture toughness associated with that law. These laws also control the evolution of one damage parameter. Various softening materials will need two or more softening laws to model all possible types of damage.

Softening laws are currently only available in OSParticulas.

Normalized Softening Law

The traction during softening is given by σf(δ) where f(δ) is the normalized softening law and σ is some scaling stress provided by the softening material. The toughness associated with the law, G_c, is determined by the area under the law times a scaling factor to account for mesh size and the scaling stress σ:

      [math]\displaystyle{ {G}_c = {V_p \sigma\over A_c} \int_0^{\delta_{max}}f(\delta)d\delta }[/math]

where V_p is particle volume, A_c is area of the initiated crack within the particle domain, and δ_max is the critical cracking strain for failure (of for when the traction drops to zero). To define a scaling law, we define a scaling factor

      [math]\displaystyle{ s = {A_c \over V_p \sigma} }[/math]

The softening material calculates s whenever softening law response is needed. All the softening law needs to evaulate is the normalized function value, f(δ,s), and the area under that law up to δ minus area under a linear return to the origin or:

      [math]\displaystyle{ A(\delta,s) = \int_0^{\delta}f(\delta,s)d\delta - {1\over2}\delta f(\delta,s) }[/math]

Each softening law requires G_c and possibly other parameters to define the law's functional form. The laws do not need to be provided δ_max because it can be calculated from the law shape, s, and G_c from total area under the law. Furthermore, δ_max depends on mesh size while G_c is a material property for damage.

Law Stability

For damage propagation to always make physical sense, damage cracking strain must increase whenever the associated applied strain is increasing and is causing damage. This stability condition requires:

[math]\displaystyle{ \max\bigl(-f'(\delta,s)\bigr) \lt {1\over \varepsilon_0} }[/math]

or maximum slope of the softening law is less then one over strain to initiate that damage process. In general, the maximum slope will be proportional to particle size. Thus, it is always possible to get stable softening laws by high enough resolution (or small enough particle size). The need for high resolution is most common for brittle materials.

Softening Laws

You create a softening law using one of "Softening" properties allowed for the selected softening materials. You pick the law by name or ID. The following table lists the available softening laws.

Name	ID	Description
Linear	1	Linear softening law.
Exponential	2	Exponential softening law.
CubicStep	3	Cubic step function softening law.

Click each one for more details and information on their properties. After picking the law, all its properties are set using property commands within the softening material definition by prefacing the law property with appropriate codes. For example, to set mode I and mode II toughnesses in an isotropic softening material, the properties are entered as I-Gc and II-Gc where I- or II- preface the Gc softening law property.

Softening Law Implementation

It is easy to add new softening laws, although unclear if that change would be much different then using the softening laws all ready available. As explained in Ref. , the addition of a softening only requires code to return:

      [math]\displaystyle{ f(\delta,s), \quad f'(\delta,s), \quad A(\delta,s), \quad{\rm and}\quad\max\bigl(-f'(\delta,s)\bigr) }[/math]

Based on just these softening law properties, softening materials have to numerically solve the differential equation

      [math]\displaystyle{ {d\delta\over d\varepsilon} = {1\over 1+ \varepsilon_0 f'(\delta,s)} }[/math]

where [math]\displaystyle{ \delta }[/math] is a maximum cracking strain associated with strain increment [math]\displaystyle{ \varepsilon }[/math]. In discrete form, the task is solve for [math]\displaystyle{ d\delta }[/math] in

      [math]\displaystyle{ d\varepsilon = d\delta + \varepsilon_0 \bigl(f(\delta+d\delta,s) - f(\delta,s)\bigr) }[/math]

The generic softening law code solves this equation using Newton's method with bracketing. The initial brackets are:

      [math]\displaystyle{ d\varepsilon \le d\delta \le d\varepsilon + \varepsilon_0f(\delta,s) }[/math]

The lower limit is found because the second term is less than or equal to zero. The upper limit is if failure occurs such that [math]\displaystyle{ f(\delta+d\delta,s)=0 }[/math]. If [math]\displaystyle{ d\delta_n }[/math] is the Newton's method sequence approaching the solution, the next increment by Newton's method is derived to be:

      [math]\displaystyle{ d\delta_{n+1} = d\delta_n - {{d\delta_n-d\varepsilon\over \varepsilon_0f(\delta,s)} + {f(\delta+d\delta_n,s)\over f(\delta,s)}-1\over {1\over \varepsilon_0f(\delta,s)} + {f'(\delta+d\delta_n,s)\over f(\delta,s)}} }[/math]

When the next increment is outside the current brackets, a custom bracketing calculation is used to make the solution always stable. Although generic code can solve this differential equation for any softening law with provided properties listed above, some softening laws may have more efficient solutions. For example Linear Softening can exactly solve the differential equation without needing Newton's method and Exponential Softening can evaluate the ratio terms slightly more efficiently then separate evaluations of the numerators and the denominators. Any new softening laws can similar override generic functions if Newton's method's or function ratios code if they have a more efficient solution.