contains the material model for the time-dependent

desirable to perform 3-D analyses on all structures,

properties for creep, shrinkage, and the aging modu-

the complexity associated with performing 3-D analy-

lus of elasticity. Other parameters used within the

ses is such that this approach is not yet practical.

UMAT subroutine are Poisson's Ratio, the cracking

Therefore, 2-D analyses of transverse strips should be

strain, the coefficient of thermal expansion, band-

used for most investigations. Additionally, behavior

width factors for creep and shrinkage, the time of set,

of the structure in the out-of-plane direction may be

and the 3-day compressive strength of the concrete

determined using 2-D strips in the longitudinal direc-

(Annex 4, Appendix A, is a complete list of input

tion and/or a 3-D analysis. The primary purpose of a

parameters required for a NISA). Curves contained

3-D analysis should be for cases where 2-D analysis

in the material model must be fitted for the concrete

is inappropriate because of geometrical configuration

mixture which is expected to be used at the project

or loading conditions.

site. This curve fitting is currently performed by

CEWES-SC, but efforts are in progress to implement

a procedure which will allow curve fitting to be

tional FE modeling techniques should be applied to

accomplished by others. The material model used in

develop an FE mesh for a NISA (ETL 1110-2-332,

UMAT was developed by ANATECH Research Corp.

"Modeling of Structures for Linear Elastic Finite

and is contained in the ANACAP-U software

Element Analysis," and Technical Report ITL-87-8,

(ANATECH Research Corp. 1992). The mathemati-

"Procedure for Static Analysis of Gravity Dams

cal relationships adopted for the material properties

Using the Finite Element Method - Phase Ia" (Will

contained in the subroutine are contained in the

1987)). In addition, consideration should be given to

ANACAP-U Theory Manual (ANATECH Research

the items discussed below.

Corp. 1992). In addition, ANACAP-U uses a

smeared crack model (discussed in further detail in

(1) General. Typically, the mesh developed for

paragraph A-5*e*(4), Appendix A) where cracking

use in the heat transfer analysis will be used in per-

occurs at the integration points. The cracking criteria

forming the stress analysis. Using the same mesh for

is based on an interaction curve between stress and

both analyses ensures that each node in the stress

strain (ANACAP-U Research Corp 1992). There is a

mesh has a temperature associated with it from the

brief description of the cracking model in Annex 2,

heat transfer mesh. In addition to items discussed in

Appendix A.

subsequent paragraphs, the designer should attempt to

include at least two elements through the thickness of

(4) Units. The ABAQUS code does not provide

any member.

an input option for different units for different param-

eters, therefore the user must ensure that all of the

(2) Foundation. For soil-founded structures, the

input parameters have consistent units. The preferred

soil elements used in the heat transfer analysis are

units are:

usually replaced by springs in the stress analysis to

save computing time. A similar approach may be

Length - inches

used for rock foundations but may require some

investigative analysis since, to date, no NISA's have

Weight - pounds

been performed on rock-founded structures. Another

possibility for rock foundations is the use of superele-

Stress - lb/in.2

ments which allows for the stiffness of a large num-

ber of discrete elements to be lumped into a single

Time - days

region which is the superelement. Despite the

method used, a rock foundation should model an area

Temperature - oF

1.5 times as deep and 3.0 times as wide as a struc-

ture's base as established in "Static Analysis of

Heat - Btu

Gravity Dams Using the Finite Element Method,

Foundation Effects - Phase Ib" (Jones and Foster in

preparation).

MCS for which a 2-D analysis of a monolith will be

(3) Inclusion of lift joints. Development of an

sufficient. However, any structure undergoing tem-

FE mesh for a NISA study must account for the

perature loadings exhibits some 3-D behavior, even

locations of the construction lift joints and vertical

traditionally 2-D type structures. While it may be

A-17