were a 2.0% maximum change in *M*r from

**Table 5. Format of a NELAPAV input file (from Irwin and**

**Speck 1986).**

one iteration to the next and a total of 10

iterations.

Subject

input parameters

NELAPAV allows the use of nonlinear

Header

problem description

(i.e., stress-dependent) modulus values in

Units

input units

the analysis. Modulus values for thawing

and unfrozen fine-grained soils are highly

Load

load, pressure, load radius

nonlinear as illustrated in Figure 5. Table

*n*

Layered system

4, from Yang (1988), illustrates the vari-

*M*1 E1 seed ν1 γ1 K0,1 h1 k1,1 k2,1 k3,1 k4,1 t1

ous types of linear and nonlinear models

*M*2 E2 seed ν2 γ2 K0,2 h2 k1,2 k2,2 k3,2 k4,2 t2

currently available for use in NELAPAV;

.

however, models 2 and 7 are not currently

.

incorporated in the rest of the CRREL

.

design procedure. In the CRREL version

*M*n En seed νn γn K0,n hn k1,n k2,n k3, n k4,n tn

of NELAPAV, model 1 has been changed

Tolerance

maximum iterations, tolerance

to the semi-log form:

*r*1, *z*1

Calculation points

*M*r =*k*1e(*k*2θ)

(10)

r2, *z*2

.

where *k*1 and *k*2 are constants and θ is bulk

.

stress. For this study, we utilized models

.

0, 1, 3, and 4, the specifics of which will

*r*m, *z*m

be described in a later section.

Notes:

One shortcoming of NELAPAV is that

*n *= number of layers

it allows only one circular load to be ap-

**Supplied for each layer**

plied at the surface. This is not a major

*M *= model number (see Table 4)

problem in this analysis because we are

*E *seed = seed modulus (lb/in.2)

dealing with roadway pavements rather

ν = Poisson's ratio

than airport pavements, which experience

γ = density (lb/ft3)

much more complex tire configurations.

*K*0 = lateral earth pressure coefficient

Table 5 is a brief listing of the informa-

*h *= thickness (in.)

tion included in a NELAPAV input file

*k*1.. *k*4 = constants to be used in models (see Table 4)

*t *= mean temperature (C)

(Irwin and Speck 1986). It may contain

up to 25 layers, with the following values

*r*m, *z*m = radius, depth at which to make calculations (in.)

for each: model number (from Table 4),

seed modulus, Poisson's ratio, total den-

During this study, *K*0, the coefficient of lateral

sity, lateral earth pressure coefficient, thickness,

earth pressure, was assigned to the layers as fol-

and the constants required for the equations from

lows: a value of 1.5 was used for paving materials

Table 4.

and frozen soil; a value of 1.0 was used for all

NELAPAV requires a seed modulus to begin

unfrozen soil layers. We now feel that this system

its predictions of modulus for the various material

is too simplified, and are modifying the section of

layers. The value used for the seed modulus is the

TRANSFORM that assigns *K*0.

resilient modulus calculated in and passed from

For each layer, NELAPAV expects to receive a

TRANSFORM. The constants required for com-

value for the total density, which is adjusted to a

putations are also passed from TRANSFORM.

The *k*2 value is a constant assigned on the basis of

buoyed density below the water table. In this study,

material type. The *k*1 constant varies with the mois-

however, we used dry density instead of total den-

ture/density level of the layer, and is calculated by

sity, and neglected to adjust to a buoyed density

multiplying together all the nonstress terms in the

below the water table. A short sensitivity study,

predictive equation for the material (Table 2).

conducted when these problems were discovered,

11