AD-774 801

MATERIAL PROPERTIES FOR POSTSHO' MIXEDCOMPANY ANALYSIS: RECOMMENDATIONS BASEDON RECENT LABORATORY AND IN SITU TESTDATA

John Q. Ehrgott, et al

Army Engineer Waterways Experiment Station

Prepared for:

Defen-se Nuclear Agency

January 1974

DISTRIBUTEL BY:

National Technk',ai Information ServiceU. S. DEPARTMW.T Aj COMMERCE5285 Port Royal Road, Springfield Va, 22151

[ - -- .. , • ' -

Unclassi fielSECURITY CLASSIFICATION OF T-.IS PAGE frhen Date EnIered)

GEFORE COMPLETING FORM

I. REPORT NUMBER 2. GOVT ACCESSION NC. 3. RECIPIENT'S CATALOG NUMBER

Miscellaneous Paper S-74-I4. TITLE (end Subtitle) TYPE OF REPORT & PERIOD COVERED

MATERIAL PROPERTIES FOR POSTSHOT MIXED COMPANY Final reportANALYSIS: RECOMEDATIONS BASED 01 RECENT i Finalre__r_LABORATORY AND IN SITU TEST DATA 6. PERFORMING ORG. REPORT NUIIBER

7. AUTHOR(&)aJ-OIRC OR GRANT NUMBER(s)

John Q. EhrgottJohn G. Jackson, Jr.

9. PERFORMING ORGANIZATION NAME AND ADDRESS 10. PF3GRAM ELEMENT. PROJECT. TASKAHEA 8 WORK UNIT NUMBERSU. S. Army Engineer Waterways Experiment Station Subtask B209

Soils and Pavements Laboratory Work SB20l

P. 0. Box 631, Vicksburg, Miss. 35180 Work Unit 11

I. CONTROLLING OFFICE NAME AND ADDRESS 12. R9.PURT DATE

Defense Nuclear Agency Janiuarvr 1974Washington, D. C. 20305 13. NUMBEROF,GES

14. MONITORING AGENCY NAME 8 ADURESS(ii different from Conti lling Office) iS. SECURITY CLASS. (of this report)

Unclassified

15a. DECL ASSIFICATION/DOWNGRADINGSCHEDULE

16. DISTRIBUTION STATEMENT (of lhis Report)

Approved for public release; distribution unlimited.

17. DISTRIBUTION STATEMENT (of the abstract enlered in Btock 20. It different IPo-m Report)Reprojuced by

NATIONAL TECHNICALINFORMATION SERVICEU S Department of Commerce

Springfield VA 22151

1,. SUPPLEMENTARY NOTES

This research was sponsored by the Defense Nuclear Agency under Subtask SB209,Work Unit 11, "Laboratory Studies of the Response of Soil and Rock to Blast-

Type Loadings."

19. KEY WORDS (Continue on reverse side if necesiory and identify by block number)

Field tests Mixed Company (Event III)Geologic materials Rock propertiesGround motion Soil propertiesLaboratory tests

20. ABSTRACT (Continue on reverse side if necessAry and Ideltify by block number)

An idealizied, horizontally layered geologic profile and an associated set ofconstitutive properties for each layer in this profila were developed duringFY 1972 for use in pretest ground motion calculations for Mixed CompanyEve t III, a 500-ton high explosive test. The study reported herein was con-dfcted during FY 1973 to evaluaie the validity of the preshot recommendationsin the light of ne data and to make appropriate recommendations for a revisedsite profile and a revised set of constitL-ive properties for use in a

FORM " 7"DDI JAN 73 47 EDITION OF I NOV 65 IS OBSOLETE Unclassified

SECURITY CLASSIFICATION OF THIS PAGE (When Dat Entered)

SUnclassi fj edSECURITY CLASSIFICATION OF" THIS PAGE(When Data Enrered)

ABSTRACT (Continued)

postshot recalculation of the event. The pretest calculational grid extendedonly to a depth of about 300 feet. Examination of possible travel paths showedthat reflection from deeper interfaces tould have arrived in time to influencethe ground motions measured during Event III. Therefore, it was recommendedthat the recalculation grid extend to depths of 500 to 600 feet, i.e., into thePrecambrian basement. Tests on samples taken just after the event showed thatwater content changes due to the wet weather conditions at shot time substan-tially affected the compressibility of the upper 2 to 3 feet of overburden soil.Thus, it was recommended that the overburden soil be idealized into two layersrather than into one as or ginally recommended for preshot calculations. Anextensive reanalysis of the available surface retraction survey data, along withanalysis of the Event III ground motion data, resulted in revisions to theentire seismic velocity profile and, hence, to the associated values of theinitial uniaxial strain modulus. The stress levels asso-iated with these moduliwere originally assumed to be quite low, e.g., less than 10 psi for the KayentaFormation materials. A lower bound value of 50 psi was recommended as beingmore reasonable for the initial Kayenta "precursor" stress. This phenomenonappears to be a function of loading rate. Tests on the Kayenta materials alsorevealed that horizontal-to-vertical anisotropy undoubtedly affected the groundmotion results, but the problem of how to utilize the various horizontal andvertical data to specify meaningful "effective" or "average" isotropic propertyvalues still remains. As a result of recent test data from Lawrence LivermoreLaboratory, however, substantial changes have been made in the recommendedfailure envelopes, which now reflect a highly nonlinear behavior, includingsignificant strength increases at high pressure.

UnclassifiedSFCURITY CLASSIFICATION OF -HIS PAGE(then Data Entered)

THE CONrVKS OF TIIIS REPORT ARE hOT TO

BE USED FOR ADVERTISING, PUBLICATIOI,

OR PROMOTIONAL PURPOSES. CITATION OF

TRADE NAMES DOES NOT CONSTITUTE AN OF-

FICIAL ENDORSEMENT OR APPROVAL OF THZ

USE OF SUCH COMMERCIAL PRODUCTS.

3

PREFACE

This paper was prcpared for presentation at tbh Defense Nuclear

Agency (DiTA) Strategic Structures Long-Range Planninag Meeting (LRPM)

held at the Stanfcrd Research Institute, Menlo Park, California., during

15 to 17 May 1973. The work described herein was conducted by the U. S.

Army Engineer Waterways Experiment Station (WES) fo: DAA under Subtask

SB209, Work Unit 11, "Laboratory Studies of the Response of Soil and

Rock to Blast-Type Loadings." Mr. C. B. McFarland was the DNA Project

Officer for Subtask SB209.

The study was performed by personnel of the 'ii7S Sils and Pavements

Laboratory (S&PL), Mr. R. W. Peterson, Soil Dynamics Division (SDD),

S&PL, conducted the laboratory tests. The analyses were performed by

Mr. J. Q. Ehrgott, who presented the results at the LRPM. Mr. J. R.

Cuxro, Jr., Earthquake Engineering and Vibrations Division, S&PL, pro-

vided helpful comments and interpretations regarding the seismic refrac-

tion survey data, and Dr. P. F. Hadala, SDD, S&PL, provided valuable

advice and guidance throughout the study. The paper was jointly pre-

pared by Mr. Ehrgott and Dr. J. G. Jackson, Jr., Chief, SDD.

BG E. D. Peixotto, CE, and COL G. H. Hilt, CE, were Directors of

WES and Mr. F. R. Brown was Technipal Director during the conduct of

this study and the preparation of this paper. Mr. J. P. Sale and

Mr. R. G. Ahlvin were Chief and Assistant Chief, S&PL, respectively.

4

CONTENTS

PREFACE ------------------------------------------ 4CONVERSION FACTORS, BRITISH TO METRIC UNITS OF MEASUREMENT----.-----. 9

CHAPTER 1 INTRODUCTION ------------------------------------------ 10

1.1 Background ------------------------------------------------ 101.2 Purposes ---------------------------------------------- 111.3 Scope ----------------------------------------------- 11

CHAPTER 2 SITE PROFILE ------------------------------------------- 12

2.1 Depths to Major Geologic Interfaces ------------------------- 122.2 Idealization of Primary Geologic Units ---------------------- 132.3 Recommended Postshot Profile ------------------------------- 15

CHAPTER 3 SEISMIC VELOCITIES ------------------------------------- 19

3.1 Reanalysis of Surface Refraction Survey Data ---------------- 193.2 Analysis of Explosive Test Data ----------------------------- 213.3 Recommended Postshot Seismic Velocities ---------------------- 23

CHAPTER 4 CONSTITUTIVE PROPERTIES --------------------------------- 30

4.1 Effect of Water Content Changes ----------------------------- 304.2 Effect of Seismic Velocity Changes --------------------------- 314.3 Effect of Horizontal-to-Vertical Anisotropy ----------------- 334.4 Recent In Situ and Laboratory Strength Data ----------------- 344.5 Recommended Postshot Constitutive Properties ---------------- 35

CHAPTER 5 CONCLUSIONS AND RECOMMENDATIONS ---------------------- 45

REFERENCES ------------------------------------------------------- 48

APPENDIX A REPRESENTATIVE STRESS-STRAIN AND STRENGTH RELATIONS ---- 50

TABLES

1 Recommended Postshot Profile and Composition Properties ----- 162 Recommended S ismic Velocity Value -------------------------- 25

A.1 Summary of Postshot Profile and Composition PrQperties forMixed Company Analyses -------------------------------------- 51

FIGURES

1 Subsurface profile showing possible paths of waves reflectedfrom deep geologic "nterfaces and corresponding arrivaltimes at the gro1i-d surface --------------------------------- 17

2 Pofile of natural soil overburden layer showing variationof water content and dry density with depth ----------------- 18

3 Profile of composite fill and natural soil overburdenlayer showing variation of water content and drydensity with depth ----------------------------------------- 18

5

FIGURES

4 Plan view of Mixed Company site showing seismic surveylocations with respect to the Event III GZ, the WES andAFWL gage lines, and the AFWL CIST experiment --------------- 26

5 Cross sections through Mixed Company site indicating seismicvelocities and estimated seismic and geologic interfaces---- 27

6 Arrival time of first motion versus range along the WESgage lines, with possible interpretations in terms ofrefracted seismic velocities -------------------------------- 28

7 Summary plot of seismic velocity profiles in Mixed Companyrocks (based on interpretations of various data) comparedwith estimated geologic profile -------------------- 29

8 Recommended UX relations for Layers 1A, 1B, and alternateLayer 1C -------------------------------------------------- 37

9 Comparison of pretest recommended UX relations forunweathered Kayenta sandstone with gun data ----------------- 38

10 Comparison of loading times versus peak stress forlaboratory tests with those estimated from fieldtest measurements ------------------------------------------ 39

11 Enlarged plots of dynamic UX test data from five specimensof unweathered Kayenta sandstone ---------------------------- 40

1.2 Results of horizontally and vertically oriented UX testson uniform Kayenta sandstone specimens and specimens withclay seams ----------------------------------------------- 41

13 Preliminary results of TX tests by LLL on virgin specimensof Mixed Company sandstone and specimens previouslysubjected to 7-kbar hydrostatic confinement --------------- 42

14 Comparison of UX loading and unloading relation recommendedfor posttest Layer 3 with relation recommended for pretestLayer III -------------------------------------------------- 43

15 Comparison of low-pressure TX failure envelope recommendedfor posttest Layer 3 with relation recommended for pretestLayer III -------------------------------------------------- 44

A.1 Representative a. versus ez relation for uniaxial strainwith unloading curves from yz = 50 and 200 psi for LayerIA -- 52

A.2 Representative az versus ez relation for uniaxial strainwith unloading curves from oz = 200, 500, and 1,150 psi forLayer A -------------------------------------------------- 53

A.3 Representative oz versus ez relation for uniaxial strainto az = 6,000 psi for Layer IA --------------------------- 54

A.4 Representative (az - Or) versus p stress paths foruniaxial strain and (oz - ar)max versus p failure

envelope for triaxial shear to p = 1,200 psi for Layer 1A-- 55A.5 Representative az versus ez relation for uniaxial strain

with unloading curves from az = 100 and 500 psi for LayerB -------------------------------------------------------- 56

A.6 Representative az versus ez relation for uniaxial strainwith unloading curves from az = 500 and 1,950 psi for Layer1B -------------------------------------------------------- 57

6

FIGITRES

A.7 Representative oz versus ez relation for uniaxial strainto oz = 30,000 psi for Layer B --------------------------- 58

A.8 Representative (u. - Or) versus p stress paths foruniaxial strain and (uz - ar)max versus p failure

envelope for triaxial shear to p = 300 psi for Layer 1B ---- 59A.9 Representative (az - ur) versus p stress path for

uniaxial strain and (u. - Or)max versus p failure

envelope for triaxial shear to p = 6,000 psi for Layer lB-- 60A.l0 Representative (az - ar) versus p stress -oath for

uniaxial strain and (az - or) versus p ;2ilure

envelope for triaxial shear to p = 20,000 psi for Layer1B ------------------------------------------------------- 61

A.1 Representative a. versus ez relation for uniaxialstrain with unloading curve from 0z = 200 psi for Layer 2-- 62

A.12 Representative 0z versus e. relation for uniaxials rain with unloading curves from az = 500 and 1,000 psii r Layer 2 ----------------------------------------------- 63

A.13 Representative 0z versus e. relation for uniaxialstrain with unloading curves from a = 500, 1,000,5,000, and 8,000 psi for Layer 2 --------------------------- 64

A.lh Representative (a. - or) versus p stress paths foruniaxial strain and (G. - or)max versus p failure

envelope for triaxial shear to p = 1,200 psi for Layer 2--- 65A.15 Representative (az - Cr) iersus p stress path for

uniaxial strain anC (a. - Orm v sus p failurea'maxenvelope for triaxial shear to p = 6,000 psi for Layer 2--- 66

A.16 Representative a. versus e. relation for uniaxialstrain with unloadiig 'curves -Prom Oz = 200, 400, and500 psi for Layer 3 --------------------------------------- 67

A.17 Representative az versr.s ez relation for uniaxialstrain with unloading :urves from a. = 500, 1,000,2,000, and 8,000 psi for Layer 3 ----------------------- 68

A.18 Representati e (cz - or) versus -) stress paths foruniaxial strain and (az - ar)max versus p failure

envelope for triaxial shear to p = 1,200 psi for Layer 3--- 69A.19 Representative (oz " 'r) versus p stress path for

uniaxial strain and (Cz - r)max versus p failure

envelope for tripxial shear to p 6,000 p.si for Layer 3--- 70A.20 Representative a. versus ez relation for uniaxial

strain with unlop ing curves from a. = 200 and 500 psi forLayer 4 -. ------------------------------------------ 71

A.21 Representative 0Y versus c7 relation for uniaxialstrain with unloading curves from cz = 4,000 and 10,000psi for Layer 4 ------------------------------------ 72

7

F: _ . . . . . . - - . . . . . . . . .. . . .. . . .

FIGURES

.1.22 Representative (az - or) versus p stress paths foruniaxial strain and (a. - ar)mlx versus p failure

envelope for triexial shear to p 1,200 psi for Layer 4-- 73A.23 Represe-ntative (az - ar) versus p stress path for

unia:xial strain and (az - ar) ax 3% rsus p failure

envelope for triaxial shear to p 6,000 psi for Layer 4-.- 74A.24 Represent:tive az versus ez relation for uniaxial

strain vAth unloading curves from cz = 200 and 500 psi forLayer 5 ------------------------------------------ 75

A.25 Representative az versus ez relation for uniaxialstrain with unloading curves from az = 2,000 and 8,000 psifor Layer 5 --------------------------------------- 76

A.26 Representative (z - ar) versus p stress paths forwuiaxial strain and (az - or)max versus p failure

envelope for triaxial shear to p = 6,000 psi for Layer 5- 77A.27 Table of elastic constants and (az - ar)max versus p

failure envelope for triaxial shear to p = 8,000 psi forLayers 6 and 7 ------------------------------------- 78

8

CONVERSION FACTORS, BRITISH TO METRIC UNITS OF MEASURE1NT

British units of measurement used in this report can b- converted to

metric units as follows:

Multiply By To Obtain

feet 0.3048 meters

tons 907.1846 kilograms

feet per second 0.3048 meters per second

pounds (mass) per cubic 16.0185 kilogrqvs per cubic meterfoot

pounds (force) per square o.6894757 newtcas per square centimeterinch

kips (force) per square o.6894757 kiloneiotonF per squareinch centimeter

pounds (force) per square o.6894757 newtons per square centimeterinch per second per second

9

CHAPTER 1

INTRODUCTION

1.1 BACKGROUND

A geologic profile for the Mixed Company site and associated con-

stitutive properties for each layer in this profile were developed dur-

ing FY 1972 by the U. S. Army Engineer Waterways Experiment Station

(WES). The idealized profile and properties (Reference 1) were dissem-

inated in June 1972 for use in developing fits to constitutive models

that were in turn used for pretest ground motion calculations (Refer-

ence 2) of Event III, a 500-ton high explosive test. Since these prop-.

erties were disseminated, additional field and laboratory test data, as

well as ground shock data from the event itself and frow the Air Force

Weapons Laboratory's (AFWL) Cylindrical In Situ Test (CIST) explosive

experiment, have become available.

Ground shock data obtained both by WES and AFWL were presented at

the Mixed Company project review meeting in March 1973. The pretest

ground motion predictions did not agree very well with either set of

measurements. The calculated arrival times for the initial ground mo-

tions were much later than those indicated by the field measurements,

and the general characteristics of the calculated wave forms did not

match those of the field data. A number of assumptions were made in the

process of developing the preshot profile and property idealizations

(Reference 3), which if not valid could well account for the observed

discrepancies between the field measurements and the calculated ground

motions. It should also be noted that the mathematical models had to be

hurriedly fit in order to meet the schedule for the pretest calculation,

and they did not closely replicate all facets of the recommended consti-

tutive properties. Also, the calculational grid did not extend deep

1A table of factors for converting British units of measurement tometric units is presented on page 9.

10

enough to incorporate the reflecting interface between the Triassic sed-

imentary materials and the Precambrian basement rock. Thus, there are

sufficient uncertainties to warrant: (1) a reevaluation of the profile

and properties in the light of the newly available data, and (2) a re-

calculation of Event III with a much mor6 precise fit to the constitu-

tive properties and profile resulting from this reevaluation.

1.2 PURPOSES

The purposes of this study were to evaluate the validity of the

preshot Mixed Company Event III site profile and the associated set of

constitutive properties in light of new data and to make appropriate

recommendations for a revised site profile and a revised set of consti-

tutive properties for use in a postshot recalcultation of the event.

1.3 SCOPE

This paper presents the results of a reevaluation of the Mixed Com-

pany site profile and constitutive properties uflizing both field and

laboratory data. Revisions to the subsurface profil were based pri-

mexily on the water contents determined in the soil overburden at shot

time. An extensive reanalysis of the seismic velocity data was con-

ducted. New constitutive property recommendations for the overburden

soil were developed based on laboratory tests conducted on specimens

taken in the field just a few days after the event. The properties of

the Kayenta Formation were revised to consider the effect of loading

rate on the initial portion of the uniaxial strain (UX) test stress-

strain curves, the effects of anisotropy on the observed triaxial shear

(TX) and UX test behaviors, and the correlation observed between shear

strength obtained from laboratory tests and strength obtained from an

in situ test.

ll1

CHAPTER 2

SITE PROFILE.

In developing the pr--cst calculation profile for the Mixed Company

Event III site, many assumptias ;ere necessary in order to translate

the available data into an icealized, horizontally layered profile. The

purpose .f this chapter is to evaluate these assumptions in the light of

additional information aad analyses and to recommend appropriate revi-

sions. As wa3 the ca,.e for the original profile, the revised profile is

intended to be an idealized representation of conditions along the WES

gage line.1

2.1 DEPTHS TO MAJOR GEOLOGIC INTERFACES

Four major geolcgic interfaces were originally defined for the site

profile. The first, at a depth of 5 feet, depicted the change from soil

overburden to the complex Kayenta Formation of intermixed siltstones,

sandstones, madstones, and conglomerates. At 70 feet, the sandstones of

the Wingate Formation vere assumed to be encountered, followed by Chinle

Formation siltstone at 400 feet. Finally, an interface with the Precam-

brian basement co, plex was set at a depth of 500 feet. Nc reason has

been found to alter these interface depths. The pretest calculational

grid, however, extended only to a depth of approximately 300 feet and

thus included only the overburden soil, the Kayenta Formation, and a

portion of the Wingtte. Since that time, several calculations have been

made for other pojects that indicate that near-surface ground motions

can be significantly affected by reflections from very deep geologic in-

terfaces. This determination prompted an examination of the profile to

see if signals from the omitted Mixed lompany interfaces were being

1 Two gage lines were installed for Event III, one by WES and one by

AFWL. A rolled fill was constructed along the WES line in an attemptto provide a uniform 5-.foot-thick layer of soil overburden; fill wasnot placed along the AFWL line, and the natural soil cover varied inthickness from 3-1/2 to 5 feet.

12

Irecorded. Figure 1 shows the subsurface profile to a range of'

1,400 feet and to a depth of 300 feet. Also shown a-:e travel paths that

are possible for some of the reflected waves and the surface arrival

times of these waves based on an assumed 7,500 ft/sec wave speed in both

the Kayenta and Wingate Formations. This approximate model shows that

signals reflected from the Chinle and Precambrian layers could have ar-

rived in time to influence the ground motions measured during Event III,

since significant near-surface motion occurred for at least 500 msec at

most of the ranges shown in the figure. It is therefore strongly recom-

mended that the grid used in any Mixed Company recalculation extend into

the Precambrian basement complex.

2.2 IDEALIZATION OF PRIMARY GEOLOGIC UNITS

As denoted above, the site profile encompasses five primary geo-

logic units, i.e., the soil overburden, Kayenta sandstones, Wingate

sandstones, Chinle siltqtonls, and the Precambrian basement rocks. In

the idealized pretest profile, the overburden soil was represented as a

single horizontal layer, Layer I. The underlying Kayenta Formation,

however, was subdivided into four layers: Layer II, representing the

weathered siltstcne; Layer III, the upper one-fourth of the unweathered

Kayenta rock; Layer IV, an artificial layer representing the softer

Kayenta material occurring '-indomly throughout the formaticn; and

Layer V, the lower three-fourths of the unweathered Kayenta rock. The

Wingate, Chinle, and Precambrian Formations were each represented by one

layer, Layers VI, VII, and VIII, respectively. In making this idealiza-

tion, it was assumed that the water content and density of the con-

structed berm would match those of the natural soil oveburden and also

that the water content and other compositiona) properties found to exist

at the time samples were extracted from the site would be the same as

those existing at the time of actual detonation.

Wet weather conditions existed at the site just prior to Event III,

and the water content of the upper 2 to 3 feet of overburden soil in-

creased from an average of about 7 percent, as determined at the time

preshot properties were developed, to about 15 percent at the time of

13

the shot. The variations in water content and dry density with depth in

the top 5 feet in a section in natural soil are shown in Figure 2.

From the surface to a depth of approximately 2 feet, the water content

at shot time appears to have been approximately 15 to 16 percent; at

depths below 3 feet, however, the available data indicate that the water

content may not have changed significantly from the original value of

about 7 percent. The variation in dry density with depth shown in Fig-

ure 2 indicates that no significant change in that composition prop-

erty had occurred since the time of the pretest investigation. Fig-

ure 3 shows the dry density and water content profile for a composite

section of fill and natural soil. The water content profile is essen-

tially the same as that shown in Figure 2; the dry density profile,

however, is quite different. The density of the fill itself appears to

be very uniform, but the upper 1/2 to 1 foot of natural soil underlying

the fill apparently increased in density due to compaction that occurred

during const,uction of the berm.

Primarily on the basis of the data shown in Figures 2 and 3, it

is recommended that the soil overburden along the main WES gage line be

idealized into two horizontal layers for postshot calculations rather

than into one as originally recommended for preshot calculations.

The postshot water content and densities of the Kayenta materials

did not differ from the pretest values sufficiently to warrant a revi-

sion in the subdivision of the Kayenta Formation. Although the wet den-

sity of the upper one-fourth of the unweathered Kayenta material was

lower by approximately 2 pcf than the pretest density, it was assumed

that the new value simply represented a better definition of the density

for that material rather than an alteration due to the blast or change

in climate. None of the clayey conglomerate material, which was arbi-

trarily placed at the 18- to 22-foot depth to form Layer IV in the pre-

test idealized profile, was encountered in any of the four 20-ft-deep

posttest borings. Therefore, the composition properties for this mate-

rial could not be reevaluated.

No additional new data were obtained in order to evaluate the pre-

shot idealizations recommended for the deep formations underlying the

14

Kayenta. However, the geologic inf'ormation, the limited dry density

data obtained on field samples taken at outcrops, and the water content

values that were previously ass'imed were reviewed. As a result, it was

again decided that the Wingate, Chinle, and Precambrian Formations

should not be subdivided but that each can be adequately depicted as a

single idealized horizontal layer for calculational purposes.

2.3 RECOMMeNDED POSTSHOT PROFILE

The recommended postshot profile and composition properties are

summarized in Table 1, which lists the numbers of the old preshot

layers and those of the new postshot layers, the material description,

the depth range for each layer, and the values of wet density, water

content, and the volume of air now judged to most narly represent con-

ditions along the WES gage line. Thz:e are two significant differences

between this profile and the pretest profile. First, the soil overbur-

den has been subdivided into two layers, Layers 1A and lB. However, if

the calculational grid cannot be zoned fine enough to effect this recom-

mended subdivision, then a composite layer, Layer IC, has also been de-

fined to represent the total zone of fill and natural soil overburden.

Second, it is recommended that the artificial Layer IV, formerly

located between a depth of 18 and 22 feet, be eliminated from the post-

shot calculation profile, at least initially. With the uncertai.ties

surrounding the definition of seismic velocities and constitutive prop-

erties, as will be discussed later, it appears that the most practical

approach would be to keep the profile as simrle as possible for the

planned recalculation. However, the basic .ssumption of lumping small

zones of material, which cannot be incor) rated in the profile due to

computational grid-size limitations, into one or more large layers

should be evaluated in future parametric studies.

15

ri a\ UN o _r HO

o " P4 -I 0. H H- Hl

43)

piQ4 tf\ 0)0 U\ U,\LA\0

~P, LA\ t-ICJ CA 1- I

4-3 H C \ o0 0 0. CO43) vi C. .

U~2) 0 H\D' t - \ t- -6 Cd0 P, C'.) i r- H H : -.I 0J LA \0(I) 1 q q i H H H H Hr

4-3P 0.1

0~ t

H 1 --

P4

0

E-1 . 0

H wi W d'

0 0 0 -P 43)cI) Q)-a

0 P4 0r t3 0o' 0

o 0 ci)fj ) d 4-6 'd '

03rz 4

dr P4: I

434

030

E-4 0 -i 03H0- 00) rd~r* 4

rd P 43 0)*-. cC!) 0*,-i ;. Hq M 0 )E-i 03 U) Q) 30U) 0 0 CI 0 H0 HH 1 43 R43 c)0 .P-1 H-i p3 t_ P, r 0)

v0 0 0 ai) P 0 *rl h9 1.

- 4-30

0 4-C'j(j () -:r L

P4

Q ) 0 HH H H H H~ ,-I H :> H H

4c) >

16

I

200-

uj - *

I0 --- -- "-- -

S100

RANGE, FTtzI15 MSEC t z225 MSEC

RANGE, FTGZ 500 1,000 1,400

700

T i

500 -

Figure 1 Subsurface profile showing possible paths of wavesreflected from deep geologic interfaces and correspondingarrival times at the ground surface.

17

A'

WATER CONTENT, PCT DRY DENSITY, PCF5 10 15 90 00 110 120

,,,.5_ NAURAL SOIL

5 SOOIL-ROCK INTERFACE

Figure P ?rofile of natural soil overburden layershoe vng variation of water content and dry densitywith depth.

WATER CONTENT,PCT DRY DENSITY, PCF5 10 15 90 100 110 120LI I L I '-

F- 2' FILLU-

Uj ' NATURAL SOI L

4L SOIL-ROCK INTERFACE

Figure 3 Irofile of composite ''ill and natural soiloverburden layer showing variation of water contentand dry density with depth.

]8

CHAPTER 3

SEISMIC VELOCITIES

If the seismic velocities used with the calculation profile do not

match those of the real site profile, then calculated arrival times ob-

viously cannot agree with field-measured data. Neither can the calcu-

lated wave forms match the field wave forms, since the calculated -wave

arrivals will be incorrectly phased. For this reason, an important key

to an adequate calculation of Mixed Company Event III is the correct

definition of seismic velocities. The analyses originally applied to

the seismic velocity data from 22 surface refraction lines are docu-

mented in the summary report (Reference 1); however, these data, along

with those available from four other refraction lines surveyed near the

test site, have now been reexamined wivh respect to the overall topog-

raphy of the area. In addition, da':a from the AFWL CIST experiment have

provided new information on horizontal wave velocities, and the arrival-

time data from Event III itself have been analyzed as a refraction sur-

vey to deduce possible seismic velocity profiles. The purpose of this

chapter is to present the results of these various examinations and to

recommend seismic velocity values for the site profile developed in

Chapter 2 for postshot calculations.

3.1 REANALYSIS OF SURFACE REFRACTION SURVEY DATA

A jilan view of the Mixed Company site showing seismic survey loca-

tions with respect to the Event III ground zero (GZ), the WES and AFWL

gage lines, and the AFWL CIST experiment is shown in Figure I. Seis-

mic velocity data from 21 hammer-source surface refraction lines, each

150 feet long with 15-foot geophone spacings, and one 570-foot-long ex-

plosive line wcre obtained from the :immediate vicinity of GZ (Refer-

ence 4). The short-line (150-foot-long) data indicated velocities rang-

ing from 2,100 to 3,200 ft/sec in the iw,?er 9 feet (i.e., a coafination

of soil overburden and weathered Kaye its siltstone) and velocities rang-

ing froii, 6,000 to 8,800 ft/sec in the immediately underlying unweathered

Kayenta rock. In the pretest analysis, a value of 1,800 fG/sec was

19

estimated for the overburden soil, a value of 2,500 ft/sec for the

weathered Kayenta, and values of 7,100 ft/sec for the upper one-fourth

of the unweathered Kayenta and 7,500 ft/sec for the lower three-f'. urths.The long-line (570-foot-long) survey indicated a zone of

14 ,000-ft/sec naterial at a depth of about100 feet, but this value was

not recommended for preshot calculations since it was believed to be too

high to be representative of the low-density Wingate sandstone expected

at that depth; a pretest value of 7,4oo ft/sec was recommended. None

of the refraction lines were long enough to proride seismic v,"ocities

for the Chinle siltstone and the Iecambrian basement; values of 9,000

and 15,000 ft/sec, respectively, were estimated for these final two

layers of the pretest profile.

Seismic surveys, each consisting of 25-, 50-, and 600-foot-long

refractions lines, were also run at the four locations shown in Figure 4

that lie between the main WES gage line and the northwest-trending mesa.

The data from these surveys were not considered in the initial analysis;

however, they were subsequently examined since the short lines - re run

with 5-foot geophone spacings that enabled the seismic velocities in the

overburden soil (ranging from 1,0".. to 1,400 ft/sec with an ave-age of

1,100 ft/sec) to be distinguished from those in the weathered siltstone

(ranging from 1,800 to 5,000 ft/sec with an average of 3,300 ft/s3c).

Velocities in the deeper, more competent Kayenta niateria_ seemed

to be related to the areal topography. Four cross sections through the

site are depicted in Figure 5. The geologic formation interfaces were

estimated from visual observations of the various mesa and canyon walls;

the Kayenta-Wingate interface appeared to roughly follow the surface

slope. The available seismic survey data have "aeen projected onto two

of the cross sections (Sections C-C and D-D); the lower velocity zones

(4,000 t'- 6,000 ft/sec) of the upper Kayenta material appeared to be

confined to tle lower elevation drainage patterns, while higher veloc-

ities ((,00 to 8,000 ft/sec) indicative of a more resistant material

were reco',ded in the higher elevation area around GZ. Velocities for

the lower Kayenta materials in either case appeared to be on the order

of 8,000 to 10,000 ft/sec. As shoi, in Figure 4, the WES gage lines

20

are located entirely along a knoll where 7,000-- to 8,000-ft/sec veloc-

ities should typify the upper unweathered Kayenta material. Along the

AIWL gage line, however, two profiles may exis'c. Close to GZ

(<300 feet), the profile should be similar to that under the WES gage

line, but farther out as the terraii. drops off the upper Kayenta veloc-

icies may be on the order of 4,000 to 6,000 ft/sec.

Although the four early 600-foot line surveys easily penetrated the

Wingate formation, none detected a velocity increase over the lower

Kayenta. The Wingate could have had a lower seismic ve]ocity than the

Kayenta, but this difference could onlj have been detected with an up-

hole rather than a refraction-type field survey. Seismic velocity mea-

surements made in the laboratory on samples of Wingate sandstone were in

fact lower than similar measurements made on the overlying Kayenta sam-

ples. The four lines were too shorL, however, to penetrate the C1inle

and Precambrian Formations; however, laboratory tests have indicated

that seismic velocities of approximately 11,000 vnd 18,000 ft/sec, re-

spectively, should be expected.

3.2 ANALYSIS OF EXPLOSIVE TEST DATA

On 28 Septeiber 1972, AFWL Londucted a 35-foot-deep CIST experiment

about 1 mile to the southeast of the Mixed Company Event III GZ (see

Figure 4.). The first arrival-time data indicated horizontal seismic

velocities in the upper unweathered Kayenta material ranging from about

10,000 ft/sec to about 20,000° ft/sec (Reference 5). These velocities

were much higher than those indicated by the surface refraction surveys.

Since they were measured along a horizontal plane, they may indicate a

significant horizontal-to-vertical anisotropy for the Kayenta material.

Even so, 20,000-ft/sec velocities are generally associated with sound

igneous rocks, such as granite, rather than with relatively soft sedimen-

tary rocks such as siltstone and sandstone. A cross-hole seismic survey

at the Mixed Company site is definitely recommended to assist in resolv-

ing this uncertainty. Since the CIST measurements were obtained at

higher elevations than any of the other r.asurements, it may also be

21

possible that the velocity differences are due to material and/or geo-

logic differences.

Event Ill was detonated on 13 November 1972. ',hock-front diagrams

were constructed by both WES (Reference 6) and .IFWL (Refer.nce 7) based

on the arrival ties recorded by their respective motion gages. These

diagrams indi-ated that the soil uverburden had an average seismic veloc-

ity of about 1,300 ft/sec; this agrees quite well with the 1,100-ft/sec

velocity value determined by the refraction surveys. '..he datc indicate

velocities on the order of 3,500 ft/sec for the weathered siltstone

stratum; again, this is in excellent agreement with the 3,300-ft/sec

average determined from the reraction data. The data also indicate

velocities for the unweathered Kay-nta materials in the 8,000- to

10,000-ft/sec ra:,7e which ar( also consistent with the refraction

survey results.

The only gages placed deep enough to be in the Wingate sandstone

were those directly under GZ. It is possible that, in this region where

the travel paths es.-entially paralleled the gage columns, the grout used

to fill the gage cclumns may hare influenced the arrival times, since it

had a higher P-wave velocity (i.e., aboxL- 10,000 ft/sec) than the over-

all mass of natu-ral materials. However, the shock-front diagram data

are insuffivien; to permit a definite conclusion in this regard.

The near-surface (1-1/2-foot-deep) motion gage results, however,

can be analyzed as a long-line, surface refraction survey to extract

some seismic velocity information for the Wingate and deeper forma-1

tions. One such analysis of the data obtained along'the WES gage l.nes

(LN302 and LN311W) was reported by Ballard and Leach (Reference 8). it

indicated a single 570-foot-thick stratum of 8,300-ft/sec material over-

lying 18 ,000-ft/sec material (see Example A in Figure 6). Other in-

terpretations or fits to the data are possible, as indicated by

1 This technique will not provide valid information for the materials

above a depth of about 60 to 70 feet since the airblast-dominatedrange of superseismic motion is the "effective" geophone spacing forthe survey.

22

Examples B, C, and D in the insert in Figure (. When the layer veloc-

ities do not monotonically increase with dcpt, as is the case with Ex-

amples B and C, the layer interface de,)ths cari,'t be .-alculated from the

refraction equations; therefore, only a iagh esuimate is possible.

3.3 RECOMMENDED POSTSHOT SEISMIC VELOCITIES

The shock-fr-ont or arrival-time eLiagrams from Event III and the re-

cently analyzed short-line refraction surveys both indicate that the

1,800-ft/sec seismic velocity originally recommended for the soil over-

burden was too high; a value of 1,300 ft/sec :s now recommended for

postshot calculation models. Conversely, the arrival-time data and the

short-line refraction data indicated that the 2,500-ft/sec velocity pre-

viously recommended for the immediately underlying weathered siltstone

was too low; a value of 3,500 ft/sec is now recommended.

Reanalysis of all of the available surface refraction data indi-

cates that velocities in the unweathered Kayenta materials depend on the

areal topography, with value .: within the ranges shown in Figure 7.

Along the WES gage line, a specific value of 8,000 ft/sec is now recom-

mended for the upper one-fourth of the unweathered Kayenta, and

10,000 ft/sec is recommended for the lower three-fourths. Previously

recommended values were 7,100 and 7,500 ft/sec, respectively.

As shown in Figure 7, a variely of field values were obtained

for depths associated with the Wingate sandstone, but all exceeded

8,000 ft/sec; laboratory tests, however, gave values consistently lower

than 8,000 ft/sec. A velocity of 8,000 ft/sec is now recommended in

lieu of the previous value of 7,400 ft/sec. Based solely on laboratory

test data, Chinle siltstone velocities are now believed to be on the

order of 11,000 rather than 9,0J0 ft/sec as originally recommended.

Finally, the Precambrian basement velocity is now believed to be at

least 18,000 ft/sec rather than 15,000 ft/sec. This conclusion was

derived from both the Event III arrival-time analysis and the recent

laboratory test results.

These current best estimates of the seismic velocities associated

with each of the layers in the idealized geologic profile beneath the

23

WES gage line are summarized in Table 2. Previously recommended values

are rAso tlabed for comparison.

24

01rd0) 0)

0fmrta 0 00 00 0 H 0 H 0~ 0 0

m 0 H HH

.1- r- 4 V1 c c 0 H 0

00

0 0000 00 00 0S0 0~ 00 00 00 0

.,j *-ri 0 co c0 LA\H tf\. 0 0

4-) C.i H-

010

4-3 (' l0

0

0043n

('1 A0 U)

or I?0 01 0 L O 4

0 0

*d d Id-

Cl)) a)3~

Cd 0 Q)C~

04 0 U -Pz Q) a

H 43) 0 r.CH 4343Q - 0 c

o l00)f 0. r-i M0)

U) 0) D p 'i i

43 ri 4- cfi p)j0o (D ! 0 0)

H~ ~~~ 43 0 C 0 (Y) _) -P0T'

c'S 4 0 ~ U 435

,/

N

-I BATTLESHIP 0, ROCK

" , .I-N-

NbN

WEES

LN31I W

WES

0 / \LN302/'

GZ AFWL

" LN302a

,\ .,

CIST 2/LEGEND

- - - -SEISMIC SURVEY LINES------ DRAINAGE DITCHES

SCALE IN FEET r /

2000 0 2COO

Figure it Plan view of Mixed Company site showing seismic surveylocations with respect to the Event III GZ, the WES and AFWL gagelines, and the jia,'IL CIST experiment.

26

A

6,70 -

IZ36,600 a-

6500- KAYENTA6,500 "" " -.. ,.....(Tt K) .,,

U.

0. 6.400 WINGATE

wU M W)-LJ

6,30 \

6,200 \

6,1I I Ij0 2,000 4,000 6,000 8,000 10,00D 12,000 14,000 16,000 18,000

DISTANCE, FT

SECTION A-A

J

6,450 -k6,600D

, 6,400 Lc'-,, T~~K . 7.8 . 46.'

6,350 '- T 6,5001 8.6 810

<T K 8t... T

6 , 3 0 0 - W t ' ,

6,250 1 6 4 0L I t2,000 4,000 6,000O 2,00 4,000 6,000

DISTANCE, FT DISTANCE, FTSECTION BB SECTION 0-0

6,550 -

LEGEND

, 6,500 Sr 8 SEISMIC VELOCITY, 103 FT'SEC

S0.10 . SEISMIC INTERFACE<6,450 F T K - - ESTIMATED FORMATION I'IfRFACE

Je rNTRADA FORMATION

w 6,400 W -. T K K NYENTA FORMATIONL W WINGATE FORMATION6,350 o.-- -- _..I L J

2,000 4,000 6,000DIST AtCE, FT

SECTION C-C

Figure 5 Cross sections through Mixed Company site (seeFigure h) indicating seismic velocities and estimated

seismic and geologic interfaces.

27

C) 0 0

0.0

-ciCL 43 *

0

x0 0d

w0

w 04

Wc) 0: 0W>0 Li :

02 "4

InED 0 4-)

U- 0 Cf- ,cdr

> Hd r4

I4-

o 0< 0 6

XuS '3L1VaI

28 r

u)F4E-H L

0I

4-)LU w 0

>- I F0 14 <

0~

00

<

0J 0*-I

c~<

0 00

U- 2ILUL 1citD

TF0 0~~~t/)L,

00

0 mWzu < z0 m I 0 ~4-)_ L U LU (RI0 0

z910

a.0 0 0u) - 0 0u 0q 0f- :E

Id 'H~LLu- 0. o .v-iw _j LuLU 0 a

UJ > 29

CHAPTER 4

CONSTITUTIVE PROPERTIES

The constitutive or mechanical properties recommend-d for use in

preshot ground motion predictions were selected so as to ,'e compatible

with other assumed properties, such as water content and seismic veloc-

ity. Obviously now, if the original assumptions regarding water con-

tents and seismic velocities have been modified, the associated

stress-strain and shear strength relationships must be modified. The

mathematical constitutive model formulations employed in the preshot

calculations were based on the assumption of loading-rate-independent

and isotropic material behavior. It was also assumed that the labora-

tory test results were representative of in situ properties, and, in

some cases, extrapolations of the available data were necessary in order

to specify behavior at pressures above the range of laboratory test

values. The purposes of this chapter, therefore, are to examine new

data that bear on these assumptions and suggest possible modificatios

to the preshot property recommendations.

4.1 EFFECT OF WATER CONTENT CHANGES

At the time of Event III, the surface soil at the site was influ-

enced by wet weather and construction conditions. As a result, it was

recommended in Chapter 2 that the 5-foot-thick soil overburden leyer

originally specified be subdivided into two layers. The water content

of the upper soil layer was estimated to be approximately 15 percent,

while that of the lower layer was estimated to be about 7 percent. A

series of UX tests were conducted on both undisturbed and remolded spec-

imens obtained from the site on the day after the event. As expected

(Reference 9), those specimens with 15 percent water content were ini-

tially more compressible than those with 7 p, "cent water content, i.e.,

their initial UX response was probably dominatec r intergranular fric-

tion, which increases as water content decreases. On the other hand,

the specimens with increased water content contained fewer air voids

than did the drier specimens, which caused them to stiffen or "lock up"

30

at smaller strains as these voids closed under the applied stresses.

Figure 8 shows a plot of the recommended UX relations developed for

the two soil layers (Layers 1A and 1B); an alternate layer (Layer 1C)

for the composite 0- to 5-foot depth is also given for use in the event

that the calculational grid cannot be designed to accept the two-layer

representation.

The TX failure envelopes were significantly affected by the changes

in water content. Test data from remolded specimens with 15 percent

water content gave failure envelopes that achieved limiting or maximum

principal stress difference values on the order of 300 to 500 psi under

mean normal stresses greater than about 500 psi. The failure envelopes

derived from tests on specimens with only 7 percent water content indi-

cated a 6,500-psi limiting value for applied pressures on the order of a

kilobar or greater. New yield surfaces and UX stress paths were devel-

oped for the two soil layers as a result of these changes in the shear

strength profile.

4.2 EFFECT OF SEISMIC VELOCITY CHANGES

The initial uniaxial strain modulus M. is assumed tc be directly1

related to the seismic P-wave velocity Vp and the mass density p in2the relation M. = pVp . Values of M. were thus recalculated for all

of the site materials using the revised seismic velocity and density

values given in Tables 2 and 1, respectiv.ly. Because all of the dy-

naLic UX test data indicated relatively low or soft initial moduli, it

was assumed that the extremely stiff seismic moduli applied only for

very low stress levels that were below the resolution of the UX test

measurements, e.g., less than 10 psi for the Kayenta Formation materials.

Although analyses of the data are still incomplete, the general consen-

sus at the Mixed Company Project Review Meeting, held in March 1973, was

that the maximum amplitudes of signals traveling with seismic velocity

through this formation in both the AFWL CIST experiment and Event III

were significantly larger, i.e., on the order of 200 psi.

Whereas the data from dynamic UX tests on unweathered Kayenta

sandstones did not reveal a significant region of stiff, seismic

31

velocity-associated moduli, data from high-prezaure (5-kbar and above)

gun tests conducted by both WES and the Stanford Research Institute (SRI)

certainly appear to indicate, that moduli of this magnitude extend to

stress levels very much above 200 psi. A comparison of the data from

the two types of tests is shown in Figure 9. Assuming that both of

these sets of laboratory data as well as both sets of field data are cor-

rect, one possible explanation for the wide variation in the stress

levels associated with th stiff initial loading moduli is that this

seismic "precursor" variance is due to loading-rate effects. Figure 10

shows a comparison of the loading times versus peak stress for labora-

tory static UX tests, dynamic UX tests, and high-pressure gun tests with

those estimated from the CIST and Event III field data. The order of

the data is certainly consistent with the loading-rate hypothesis.

To examine the effect of loading rate on initial uniaxial strain

response, plots of stress/strain and stress/time from five dynamic UX

tests on similar unweathered Kayenta material were enlarged as shown in

Figure 11. Three of the five specimens had a stiffer initial response

than the other two, but at axial stresses above about 20 psi, the con-

strained moduli from all five spe-iriens were essentially the same (Fig-

ure lla). Figure llb shows that t',e stress pulses applied to these

specimens had different initial loading rates, i.e., 2,000 to

3,000 psi/eie,_' for Pulse 1 as opposed to 6,000 to 7,000 psi/sec for

Pulse 2" at 20 psi, the loading rates for both pulse typ's increased to

about 30,000 psi/sec. Above 20 psi, the rates appeared to continue to

increase identically. The data in Figure 11 are admittedly too limited

to warrant a positive confirmation, but like the data in Figure '0,

they _:r'ainly tend to corroborate the loading-rate hypothEis,

Under field loadings, stresses in the superseismic region probably

rise more uniformly to peak intensity, i.e., there is no slow lead-in.

Therefore, because of the relatively slow lead-in characteristics of the

laboratory loading pulses, the U(X test specimens may not have been

loaded fast enough initially to be representi.tive of field conditions,

and hence, the seismic-associated stress le, el:s assumed for the pretest

property recommendations would have been too low. From the above

32

reexamination of the available laboratory data, it now appears that a

lower bound value of 50 psi is more reasonable for the stress level as-

sociated with seismic velocity. When the properties derived from the

AFVW: nalysis of the CIST field data become available, they may give a

better indication of the actual maximum stress level associated with the

seismic speeds. However, it is possible that CIST-derived data might be

an upper limit, since the loading rates in that experiment are co-nsider-

ably faster than those in Event III, except perhaps in the immedLate

vicinity of GZ.

4.3 EFFECT OF HORIZONTAL-TO-VERTICAL ANISOTROPY

it was obvious from the presence of bedding planes that response ol

the Kayenta rocks would be governed to some degree b. anisotropic prop-

erties. Data from direct pull and Brazil tension tests confirmed that

these materials have a substantially higher tensile strength in the

horizontal direction than in the vertical direction (Reference 1). Un-

confined compression (UC) tests were conducted by Terra Tek, Inc. (TT),

on core with bedding planes oriented parallel and perpendicular to the

loading axis (Reference 10). Strengths from the parallel oriented tests

were as much as 50 percent lower than those obtained from perpendicu-

larly oriented tests.

Recent compressibility data from VES UX tests on Kayenta sandstone

specimens. -me of which contained thin clay seams or beddings and some

oi which were uniform, are shown in Figure 12. As expected, the uni-

form specimens were less compressible; their responses to horizontal or

vertical loadings were essentially the same. On the other hand, verti-

cally oriented specimens containing horizontal clay seams were much more

compressible than the horizontally oriented test specimens. Since the

tests were limited to the relatively competent unweathered sandstones of

the Kayenia formation, nothing is known as yet about the anisotropic

compressibility characteristics of other materials, such as the weath-

ered siltstone.

As previously indicated, the codes used in the p-etest ground

motion prediction calculations employed isotropic material models.

33

Therefore, even though the laboratory data obtained for the Mixed Com-

pany sandstones definitely indicate anisotropic stress-strain and ten-

sile strength characteristics, a basic problem still remains as to how

to utilize this information in specifying "effective" isotropic property

%railz for subsequent calculations.

4.4 RECLST IN SITU AND LABORATORY STRENGTH DATA

One of the pretest analysis assumptions was that the then-available

laboratory test data on Kayenta rock specimens were representative of in

situ properties, since the Rock Quality Index of the core was greater

than 90 percent and no other data existed that could be used to adjust

the laboratory data to reflect in situ conditions. More recently, hcw-

ever, a UC tezt was conducted by TT at a rock outcrop near the CIST ex-

periment on a large block of in-place sandstone (Reference 10). Corres-

ponding laboratory tests were also conducted on this sandstone. The

field strength was aproximately 30 percent lower than the strength from

similarly oriented, small-size laboratory specimens of the same sand-

stone. In the absence of comparative data on confined specimens, it is

assumed that the unconfined results represent an upper-bound reduction

factor, i.e., in situ and laboratory test data should agree best when

the materials are confined under large static overburden or dynamic live

loadings.

Additional information supplied by Lawrence Livermore Laboratory

(LLL) on one type of Mixed Company sandstone has direct bearing on the

validity of the pretest failure envelope shapes at high pressures (Ref-

erence 11). Prior to the LLL tests, the only failure data available

were the WES low-pressure results and those obtained by TT at ionfining

pressures of 2 and 4 kbars (see solid circles in Figure 13). A nearly

linear relationship was assumed as shown in the figure in order to ex-

tend the envelope beyond that range of pressure. The LLL data confirmed

the earlier WES and TT results but not the assumed high-pressure enve-

lope. Rather, they showed an extremely nonlinear envelope that sub-

stantially increased under confining pressures in excess of 4 kbars

(see open circles in Figure 13).

34

LLL also determined the failure strengths of specimens that had

been first hydrostatically loaded to 7 kbars and then unloaded to lower

confining pressures prior to application of shearing stresses. Struc-

tural collapse apparently occurred under the 7-kbar preloading since

the resulting failure envelope was nearly linear, similar to that ex-

pected for a dense :and (see crosses in Figure 13). Thus, the sand-

stone material could have two failure envelopes, depending on its load-

ing history. Although such crushing could only occur under very intense

pressures of the magnitude expected at ranges very close to GZ, it could

affect the characteristics of wave forms propagating out from that

region.

4.5 RECOMMENDED POSTSHOT CONSTITUTIVE PROPERTIES

The complete set of representative stress-strain and strength rela-

tions for Layers 1A through 7 is presented in Appendix A. They consist

of UX axial stress-axial strain relations, UX stress paths for both

loadirg and unloading, and TX failure envelopes. The recommended post-

shot profile and composition properties were presented in Table 1, and

the revised seismic velocity values were given in Table 2.

As stated in Section 4.2, initial UX compression moduli were recal-

culated for all of the site materials using the revised seismic velocity

values. Each of the UX relations for Layers 1A through 7 reflects these

moduli (Appendix A). The constitutive properties for the two overburden

layers reflect the effect of increased water in the surface layer;

Layer 1A is initially more compressible, locks at a smaller axial strain,

and has less strength than Layer lB.

For the Kayenta materials (Layers 2, 3, and 4), the initial por-

tions of the originally recommended UX curves were altered to reflect a

50-psi stress level associated with their seismic velocities. The new

curves were then smoothly transitioned so as to merge into the previ-

ously recommended relations at a stress level of about 2,000 psi, such

as shown in Figure 14. No changes were recommended for stress levels

above 2,000 psi, primarily because of the uncertainty expressed in Sec-

tion 4.3 as to how to utilize the various horizontal and vertical data

35

to specify "effective" or "average" isotropic property values.

The un¢onfined strength data for Layer 2 were reduced by 30 percent

from the pretest envelope to account for differences between in situ and

laboratory test conditions. The envelope was gradually merged back to-

ward the values of the pretest envelope at higher mean normal stress

levels. The principal stress difference levels associated with uncon-

fined strength for Layers 3 and 4 were reduced by 20 percent for the

same reason, and the yield surfaces were then merged back into the

values of the pretest envelopes. As shown in Figure 15, such changes

are almost trivial for envelopes depicting strong Coulomb-type initial

response. More substantial changes were made at high stress levels

(20 to 60 ksi mean normal stress) where the envelopes reflect the non-

linear shape previously observed by LLL (see Figure 13).

The posttest constitutive properties for to'e Wingate Formation,

Layer 5, are the same as the pretest properties wit!, the exception that

the initial portion of the UX curve has been modified to oeflect the

50-psi seismic precursor. The remaining Chinle and Precambrian mate-

rials (Layers 6 and 7) are assumed to be elastic, with both sets of prop-

erties reflecting the revised compression wave velocities of 11,000 and

18,000 ft/sec, respectively.

36

4I

(PIETEsr LAYER r)I-AKER /B

2.5-5'

LAYER I

3 -00-2.5'--

IIII

U)' AL TERNIA TEI" LAYER/C

b" 1 0-5'

U) I-- IxI

4

0000 10 20 30

AXIAL STRAIN E Z , PCT

Figure 8 Recommended UX relations for Layers IA, 1B, -1.d

alternate Layer 1C.

37

co2

000

PA

CH

(44)

~LV4 cr

C.,)a) - :

383

~, 0xi

L43U:D3

00

4-)

E3 0

0 cf 0

U, Cd

w

00 U-

I0 CL ~

I0 p043)

0)

0 0)

*,- 0

00

ISMi SS381S M4V3d w

39

C))0 4

U) 0

;3 -) .

cajLi 4- 0 0

cql

4-'

0404' *-

0 HC~4-)

C~i

o C) CD'% 0)4-)

w\ U 4]

0~ EnX2

-0)

Cd P,

~L0 H 0)0 0

I- .I HU

0 < p ., *rl

00 0) 04) 00 0O (D .

04 *Cd

1% 1 SSO38WS1VXV

4o

TYPE LOADINGTEST SANDSTONE ORIENTATION RATE

I S UNIFORM HORIZ STATIC2S UNIFORM VERT STATIC3D CLAY SEAMS HORIZ DYNAMIC4D CLAY SEAMS HORIZ DYNAMIC5D CLAY SEAMS VERT DYNAMIC6D CLAY SEAMS VERT DYNAMIC

8- 1/7-

I/L.- 3 D6- 2S

5- 4D

II IJ -

3-

2 /

00 0.2 0.4 0.6 0.8 1.0 1.2

AXIAL STRAIN C., PCT

Figure 12 Results of horizortally and vertically oriented UX testson uniform Kayenta sandstone specimens and specimens with clay seams.

41

C> 0

0 40

00 cc 0a

C, t

E Ul co~Li 0

LI

V4J(Ouf (U)

X j

Q 4

,I

42~-

Lii YIELD

LU

L 7UX STRESS PATH

0.4 0.

MEAN NORMAL STRESS, KSISTRESS PATH

Ile

x 0 0.2 0.4 0.6 0.8AXIAL STRAIN C) PCT

UX RELATION

Figure 1~4 Comparison of' lX loading and unloading relation recommendedfor posttIest Layer 3 (solid lines) with relation recommended for pre-test Layer III (dashed lines).

43i

7

6-/

5-

b,

b/W/

U PRETESTz 4 -WJ UC

L.

POSTTESTWJ UC

z

2-

UC LOADING PATH

I-

//

0

oI I III

0.5 1 1.5 2 25 3MEAN NORMAL STRESS P, KSI

Figure 15 Comparison of low-pressure TX failure envelope recommendedfor posttest Layer 3 (solid line) with relation recommended for pretest

Layer III (dashed line).

44

CHAPTER 5

CONCLUSIONS AND RECOMMENDATIONS

The pretest ground motion predictions did not agree very well with

the ground shock measurements made during Mixed Company Event II. How-

ever, a number of invalidated assumptions were made in the process of

developing the preshot profile aid property idealizations. in addition,

the mathematical models were hurriedly fit in order to make the pretest

calculation and did not closely replicate all features of the recom-

mended profile and properties. As a result, it was concluded that

sufficient uncertainties existed to warrant: (1) a reevaluation of the

profile and properties in the light of more recent data, and (2) a re-

calculation of Event III with a much more precise fit to the constitu-

tive properties and profile resulting from this evaluation.

The pretest calculational grid extended only to a depth of about

300 feet, i.e., into the Wingate Formation. Examination of possible

travel paths showed that signals reflected from the Chinle and Precam-

brian layers could have arrived in time to influence the ground motions

measured during Event III. It is therefore strongly recommended that

the grid used in any recalculation extend to depths of 500 tc 600 feet,

i.e., into the Precambrian basement.

Samples taken just a few days after the event showed that the wet

weather conditions existing at shot time resulted in a significant in-

crease in the water content of the upper 2 to 3 feet of the overburden

soil over that determined earlier; tests on these samples showed that

the increased water substantially affected the compressibility of these

materials. It is now recommended that the postshot calculational zoning

be set fine enough to accommodate an idealization of the overburden

soil into two horizontal layers, rather than into one as orginally rec-

ommended for preshot calculations.

With the uncertainties surrounding the definition of seismic veloc-

ities and constitutive properties, it appears that the mist practical

approach at this time would be to keep the profile for the planned re-

calculation as simple as possible. Thus, it is recommended that the

45

4-foot-thicX a.tificial layer, which was included in the preshot profile

to repvesent ;he soft clayey conglomerate material occurring randomly

throughout the Kayenta Formation, be eliminated from the postshot pro-

file. However, the basic assumption of lumping small zones of material,

which cannot be incorporated in the profile due to computational grid-

size limit ttions, into one or more larger layers should be evaluated in

future parametric jcudies.

An extensive reanalysis of the avaiiable surface refraction survey

data along with anaysis of the Event III ground motion data resulted in

revisions to the seismic velocity profile as listed in Table 2. Hori-

zontal velocities measured by AFWL in its CIST experiment could not be

reconciled with the refract.on survey data; a cross-hole seismic survey

is definitely recommended to assist in resolving this uncertainty.

Based on the data and ana-yses presented herein, new constitutive

property recommendations have been developed for each of the layers in

the proposed postshot profile (Appendix A). Values of initial UX moduli

were recalculated for all of the site iaterials using the revised seis-

mic velocity values. The stress levels associated with these moduli

were originally assumed to be quite low, e.g., less than 10 psi for the

Kayenta Formation materials. A lower bound value of 50 psi is now rec-

ommended as being more reasonable for the initial Kayenta "precursor"

stress. This phenomenon appears to be a function of loading rate and

definitely requires further study.

Tests on the Kayenta materials also revealed that horizontal.-to-

vertical anisotropy undoubtedly affected the ground motion results. But

the problem of how to utilize the various horizontal and vertical data

to specify meaningful "effective" or "average" isotropic property values

still remains as a nagging item for further research.

In the preshot analysis, yield envelopes were extended in rela-

tively simple linear fashion to pressure ranges beyond that of the pre-

test data. As a Lcesult of recent LLL data, substantial changes have

been made in the failure envelopes, which now reflect a highly nonlinear

behavior, including significant strength increases at high pressures.

Care snould be taken in fitting the material models for the

46

......... ......

proposed postshot calculation in order to insure that they reflect the

revised properties as closely aF possible. Recent developments related

to cap-type constitutive models have significantly improved their abil-

ity to mirror detailed and complex nwterial property specifications. If

after doing this the recalculated ground motions still do not reasonably

match the field data, attention should be directed toward defining prop-

erties and developing simplified models for loading-rate-dependent

and/or anisotropic materials.

47

REFERENCES

1. J. Q. Ehrgott; "Preshot Material Property Investigation for theMixed Company Site: Summary of Subsurface Exploration and LaboratoryTest Results"; paper presented at Mixed Company/Middle Gust Project Re-view Meeting, 13-15 March 1973; Santa Barlara, Calif.; Unclassified.

2. i. S. Sandier, J. P. Wright, and M. L. Baron; "Data Report,Pretest Ground Motion Calculations for the Mixed Company Event of theMiddle North Series"; Contract Report, October 1972; U. S. Army Engi-neer Waterways Experiment Station, CE, Vicksburg, Mizs.; prepared byWeidlinger Associates, Consulting Engineers, under Contract No. DACA39-72-C-0002 and DACA 39-70-C-0016; Unclassified.

3. J. Q. Ehrgott; "Preshot Material Property Investigation for theMixed Company Site: Summary of Subsurface Exploration and LaboratoryTest Results"; Miscellaneous Paper S-73-6, Tables 4.! and 4.2, October1973; U. S. Army Engineer Waterways Experiment Station, CE, Vicksburg,Miss.; Unclassified.

4. R. E. Leach; "Refraction Seismic Site Investigation at Site D,Grand Junction, Colorado"; Memorandum for Record, 4 October 1971; U. S.Army Engineer Waterways Experiment Station, CE, Vicksburg, Miss.;Unclassified.

5. Air Force Weapons Laboratory, Kirtland Air Force Base, N. Mex.;Letter to: U. S. Army Engineer Waterways Experiment Station, CE,Vicksburg, Miss.; Subject: "Transmittal of Data Package"; 19 January1973; Unclassified.

6. J. K. Ingrain; "Ground Motion and Stress, Project LN302"; paperpresented at Mixed Company/Middle Gust Project Review Meeting, 13-15March 1973; Santa Barbara, Calif.; Unclassified.

7. S. P. Chisolm, Air Force Weapons Laboratory, Kirtland Air Force'B,-se, N. Mex.; Letter to: U. S. Army Engineer Waterways ExperimentStation, CE, Vicksburg, Miss.; Subject: "Shock Front Profile, ProjectLN302a"; 26 April 1973; Unclassified.

8. R. F. Ballard, Jr., and R. E. Leach; "Project LN3llW1 StrongMotion Seismic Measurements"; paper presented at Mixed Company/MiddleGust Project Review Meeting, 13-15 March 1973; Santa Barbara, Calif.;Unclassified.

9. A. J. Hendron, Jr., M. T. Davisson, and J. F. Parola; "Effectof Degree of Saturation on Compressibility of Soils from the DefenceResearch Establishment, Suffield"; Contract Report S-69-3, April 1969;U. S. Army Engineer Waterways Experiment Station, CE, Vicksburg, Miss.;prepared by M. R. Davisson, Foundation Engineer., under Purchase OrderNo. WESBPJ-68-67; Unclassified.

10. H. R. Pratt, Terra Tek, Inc., Salt Lake City, Utah; Letterto: Air Force Weapons Laboratory, Kirtland Air Force Base, N. Mex.;

148

Subject: "Progress Report on Contract F29601-72-C-0121"; 12 December1972; Unclassified.

11. 1. C. Heard, Lawrence Livermore Laboratory, Livermore, Calif.;Letter to: \. S. Army Engineer Waterways Experiment Station, CE,Vicksburg, Miss.; 5 April 1973; Unclassified.

49

APPENDIX A

REPRESENTATIVE STRESS-STRAIN AND STRENGTH RELATIONS

An idealized profile and set of constitutive properties were devel-

oped for use in the postahot calculation of the Mixed Company Event III

test. The recommended postshot profile and composition properties are

summarized in Table A.1 which lists the layer number, the material de-

scription, the depth range for each layer, the values of wet density,

water content, and volume of air, the seismic velocity, and the tension

limit (tension cutoff) of axial stress. The constitutive properties

consist of a UX axial stress-axial strain relation, a principal stress

difference-mean normal stress path for UX, and a TX failure envelope.

The constitutive properties for Layer 1A are shown in Figures A.1 to

A.4; Layer 1B properties, in Figures A.5 to A.10; Layer 2, in Fig-

ures A.11 to A.15; Layer 3, in Figures A.16 to A.19; Layer 4, in A.a)

to A.23; Layer 5, in A.24 to A.26; and Layers 6 and 7, in Figure A.27.

All of the properties reflect the response of the materials to live

stress loadings. Tension cutoff values, including the contribution due

to overburden stress, are indicated for layer interface locations, i.e.,

top and bottom of each layer.

50

In *H *rj M~ dU) A ) In) P4

0z NdP P4 P,

0 4-) 0 + +

En el r- I I I0 Iz Il

ca a)> (j4 -I H

E-1)* I~ I~ ~ ' ~

00 P 0 0A t- C 0 LA C) I

LA 0 0 LA 0A 0 00Fi00)C LA (- C (Y') -'\ 0 0 0 0

> 044 H H-

04 4-3 H- (' LA 0 0 -0 0

1 0 <C) 04 \0i t- L t O o0~fP C\J H~ H- Hi - ~ CM L C

C 4434JI 00)J U - Y -

0m U 000 0 0m 4-) M ~

O- 0 4t L - 6 C

43i Crl A l -

P44

H 0H0r C\ M 0r LA\ 0\ . 0 0g.l Citt- - 0N

0

P4 4-)4

0 1 P1Cdn)U0 cm 0 Cd\ 0 01 z)0

0FZ 430 0 (U 3- U 3 ~ 4

U) -40)C 0~ ~~d 0

Cz 1O 000~Ca) H3C rd rd

4--1 H-) 43U a) d

0 ofF~4 0'~0 H- HJ HEnC ~ A kl

Lo P H 0ZO O (5n

240-

16-

200-

12-

160-

F 120-cj, 8-

-j

x80

CD 0

4-

40*NOTE:

TENSION

CrT= TOP OF LAYERO"T 2 . PSI BOTTOM OF LAYER

0L I0 * ~4 6

AXIAL STRAIN, EZ I PCT

Figure A.1. Representative az versus c2z relation for uniaxial strainwith unloading curves from az 50 and 200 psi for Layer 1A.

52

ls>f OO9=viN

r.

Cd0

00

Cd.

0

W r

P4

N 20

4)0

0~~ 0I~~d 0

co CD (i

~~~C 'S1dS 1V0 0 0 53

0.4

5

0.3

W4 4-

(I)

wH

0.2-

2-

0.1-0.-

0 O0 10 20

AXIAL STRAIN, E . , PCT

Figure A.3 Representative az versus cz relation for uniaxial strainto rz = 6,000 psi for Layer 1A.

.r1 0

0 43 )

o A

0 1-1

00

0 W U24

-0 W 4

co (r

(1)-I

0Z H

0jNL V- CH

low ,

0

0.-0>/)

0 0 -7.

(0d 0 P

Id 0 0 0 0 0 0

(0 it) f) (00

0 0 0 00SHJVG ()M - X

-'0--7-0 '3:)N3842UIG~ SS98LS lVdl:Nld

55

cn,<

40- 600

II500-

30-

400-

U)

C,,.

WI-I

C/ 30020-

J

x

IO O I -N O T E: E S O100- 0TENSION

II / T=-.I PSI TOP OF LAYER/ -T=-4.0 PSI BOTTOM OF LAYER

o o0 I a 3

AXIAL STRAIN, EZ, PCT

Figure A.5 Representative az versus ez relation for uniaxial strainwith unloading curves from .= 100 and 500 psi for Layer lB.

56

U)

0.4-

5.0-

0.3

4.0-

C/)mC,)

(1) 3.0 -0.2-

< ';°l2.00

UC'

1 .0 -

0 C0 10 20 30

AXIAL STRAIN, Ez , PCT

Figure A.6 Representative ;z versus sz relation for uniaxial strainwith unloading curves from .= 500 and 1,950 psi for Layer lB.

57

-~~~~~~....... .. .a .i / ...

'..i .A " .l .. 1 ! - ' P ~~]

30 C,

2.0-0

2-5

20-

LU

()1.0-

-j

x10-

0.5-

5-

0- 00 1 03

AXIAL STRAIN, E7, PCT

Figure A.7 Representative a. versus c. relation for uniaxial strainto Uz 30,000 psi for Layer 1B.

58

00

Cd $

0~04,-) CH~

0o -1or (1)

P0)

U)4-

Z-

U-0

oo (1

Z 0)

z 0H

o PA

0 0

Cd

0 0 0 0tSd 0 0 0cm]

0 0 0 0 0

.0-2- '30N-33=dIG SS38-LS 1lVdl:NI~cd

59

(00

Cd

Cd0

s-4 0

4..

P, 4w

a)

U)O

oz0

-Nz -,

~0

-1-

00 -

w-o 41l)

wwM~j Ii E U)

ui 0

(0 N0

0 060

I<

V) P

0

10^

Cj

0

(I)D

4-) Ct)U)

P4*

too

0 0z P4

0ar I

Go z% "46- <

14

0(1

Hto ' Cu 0

C~j 0

-D.fl '32)N33dIO SS3?±L1S 1lVdl:NI~d

ZIrLI

U33

4f-)

b

00

0

Ho~ <

00

0,04-U 0

PII

H)

0

0 00

0 0 f

0 0 0 0 0

70 'SS3S -1IV IXV

62

0.8 12

10-

0.6-

b 8-

I-U)

6-0.4-

x

4-

0.2-

2-

AXIAL STRAIN, EZ PCT

Figure A.12 Representative az versus c.~ relation for uniaxialstrain with unloading curves from a. 5)0 and 1,000 psi for Layer 2.

63

12-

0.8-

F 100

0.6-

0

0.6-

~4 4

(I) -

w CO

oL o

66

- C i)

o Om

01

(I) rj0 t

-0 H A

0 C

0

)(L 0/207C

1- 0101---0 4-

W (D3

a0 0<

0:00 ~4

z 0

0 o

-0 Li

0P~*r

0 4.)LL0sm43a) 0

suvem r=4

9 0 0

' DD. ':N383-AdiG SSMIJS -lVdl:Nldd

65

',- 0

U') *t4 C)c

UII

00

' ~ ~ Pi (I) 4

~ '2 UitC)~) 'L gl '

0 0~

Ism0 N O c

W* W N

666

w

wJ

0 0a. 0

2I (n W

z _ _

w p~

z F-H

0

(Z 0*

"5-I

ox (npq0< 0

*H0

00

OC)HO

ISd 0000

L 0X 0

0 0 0 0

7 D 'SS3UJLS 1lVIXV67

m 12

0.8-

10-0

0.6- '

8-

W"( )

17. 6-( 0.4-

IZ00

-00

/ / /

0 0.4 0.8 1.2

AXIAL STRAIN, E ,PCT

Figure A.17 Representative az versus sz relation for uniaxialstrain with unloading curves from az = 500, 1,000, 2,000, and8,000 psi for Layer 3.

68

0

0

0 D 0 1O

4-

0 P

0

0 0" Pi

Cn10

C',,

~e. 0a. '9000<

(n47o 0

-00w N Q)

z~ 0

0

tLa. 0k

I5Xi6 1xA H

o0 0

ISd Ot-=d

cm 0

Wf- O ' :N33JIO( SSBd.LS lVdD:NIlid

69

(00

rI0

(n- 0

001) Cd "0

zl -40

0 0

)4

7 (70

>LL<0

.- 00 0

a. 0

z TT

IU% 0

0 C0

4--

U)

0 (0

0z2 P

<r4'

k0

00

SIV 0 00 00 0~~

20~~ 'SrUS1VX

Ii.d

< U)

C,)O12 -

0)0.8

0

10-

0.6

8-

(n

I-on 0.4 6-0

x-1 0

00.

4-

0.2-

1/ ____0.

.. -OT 040.8 1.2

AXIAL STRAIN, E 7 PCT

Figure A,21 Representative az versus sz relation for uniaxialstrain with unloading curves from z= 4,000 and 10,000 psi for

Layer 4.

72

U)

co

cs

-0

aC'J

d N

0 LOU)

0 w0 w 0con

P4

oo V)0

0 0)

0 wH.o0 0. r

dU)0

0 -0

Ii

-0

0~C\ 0 0C\

0 0

~.O-fl '~N~dz~dI SS3LLS VdK~I:J

73 d

(I)

0 Enn

c'I)

69-4 0

-. ~4- 1

0

0A

4.P

N Q04-'

0

I S \A

.I.0-20 ' DN383=UIG SS38.S 1lVdI)Nld

<0

IL 0

0000H

Z 4)0 C 0

z 6~w ji

00

4

<H

(I))

o X 020<

z0

(f% 0

0,4

009410

11i C\1

IsOS 8I 1 *

I I j (\i75

0.8

10-

0.6-

8-b

'Ix

- 6-0/ 0.4-

..J

x4-

ar

02 -NOTE:0.2~ TENSIONCC0T=-1 4 9 .9 PSI TOP OF LAYER

/r0=-444.4 PSI BOTTOM OF LAYER20o

0

I 0-- 0 ,0.4 0.8 1.2

AXIAL STRAIN, Ez , PCT

Figure A.25 Representative az versus . relation for uniaxialstrain with unloading curves from a = 2,000 and 8,OOo psi forLayer 5.

76

-6

i) mH

1--

0) 4-

P, c

(T)u

0.00z

)< -,9,

0

00

N a

00

o 0

0 0C~j Q)

ISpd -;V=dr

0U8 (0 00

~0-7D '3:N38I3ddIQ( SSE:I.S -lVdl:Nd

77

Co)

0

CH-

0.

I--6

zc5o

6 Z

4-'< 0)0a:

CL 7 Ii

ID C

0:-o~ 0P

co~, +m<

Is>I~~~~~L 0- -3AridOLI~ ~-dI~~~~~~. w__ _ _ _ _ _ _ _ _ _ _ _

suv4ziV 0Od. S4~-=

4rx~~~~~~~j 0 ~ ~~i S3 i V iN~78d