pne 1115 Pre-Gondola

8/6/2019 pne 1115 Pre-Gondola

1/77


2/77

PNE-1115NUCLEAR EXPLOSIONS -PEACEFULAPPLICATIONS (TID-4500)

INTERMEDIATE RANGE GROUND MOTIONSFOR PRE-GONDOLA II ANDASSOCIATED EVENTS

Dean V. PowerLawrence Radiation LaboratoryLivermore, California

April 1968

This report has been issued as UCRL-50433.


3/77

AcknowledgmentsThe author wishes to express his gratitude to all

those who contributed to this effort. In particular, theefforts of John Lane (Mechanical Engineering), KennethOlsen (Electronics Engineering), and Thomas Tami(Nuclear Cratering Group) are gratefully acknowledged.

-u-


4/77

AbstractGround motion records from seven high explosive cratering events

in northeastern Montana were analyzed for peak velocity, power spectraldensity, and velocity spectra. The events included four 20-ton singlecharges at depths of burst which varied from 42 to 57 ft, a 140-ton rowcharge consisting of three 20-ton charges and two 40-ton charges atoptimum depths of burst, and a fully coupled charge of 0.5 tons and a de-coupled charge of 0.5 tons at optimum depths of burst. Itwas found thatat these depths and charge weights an increase in depth of burst resultedin an increase in peak velocities and power spectral densities as measuredat distant points (>5 km), while no significant frequency shifts were noted.Power spectral density was found to be approximately proportional to thefirst power of yield. For this region it was determined that power spectraldensities varied inversely as radius to the 3.55 power, and peak velocitiesvaried inversely as radius to the 1.6 power. An increase in both velocitiesand power spectral densities for small decoupling factors was found tooccur for a certain explosive-cavity configuration. Three analysis tech-niques, peak velocity, velocity spectra, and power spectral density, arecompared and it is shown that power spectral density is the most consistentmethod when comparing records from different measuring stations.

-iii-iv-


5/77

ContentsACKNOWLEDGMENTSABSTRACTINTRODUCTION

Description of Project Pre-Gondola IIDescription of Events Associated with Pre-Gondola IIDescription of Other Events Related to Pre-Gondola IIObjectivesBackground

PROCEDUREStation SelectionSeismic InstrumentationRecording and TimingInformation Retrieval and ReductionData Analysis

RESULTSData Recovery and ReductionPeak Velocity and Vector AdditionComb Filter Peak VelocityPower Spectral Density

ANALYSIS OF RESULTS, DISCUSSION, ANDINTERPRETATION

iiiii-iv

1133346689

1 0111 31 31 31 32 3

Time History VariancesVelocity in the Frequency DomainPower Spectral Density and Seismic Energy

CONCLUSIONSREFERENCESAPPENDIX A. A Discussion of the Transmission of Seismic Signalsand Their MeasurementAPPENDIX B. A Discussion of Comb Filter Peak Velocity and VectorAddition Analysis TechniquesAPPENDIX C. A Discussion of the Power Spectral Density Code

565657586 16 26 36567

-v-


6/77

Figures1 Map showing environs of Pre-Gondola II and associated events. 22 Map showing location of free-field seismic stations for Pre-Gondola II140-ton row charge. (Station 2BNwas operated for Alfa, Charlie,Delta, and Bravo Events, and Stations 3N, 4N, and 5Nwereoccupied on Bravo Event during Pre-Gondola series; Stations 2BN

and 6Wwere occupied on decoupling series during Pre-Gondola II.) 73 Response of HS-I0-l velocity geophone at 35 and 70%damping. 84 Three-component instrument package during set-up and prior toinsertion into outer canister (canister is partly visible behindmetal drum on right). 95 Top view of tape recorder used for recording geophone signals(tape cartridge is shown in place at right). 106 Partly dismantled recorder, showing its size, portability, etc. 117 Peak velocities for Pre-Gondola IBravo Event as a function ofdrstance (large value at 96 km is due to single pulse of very highamplitude in velocity-time signature; rest of signature is ofrelatively low amplitude). 158 Peak velocities for Pr-e -Gondola- II 140-ton row charge as a functionof distance. 159 Comb filter peak velocity as a function of frequency for Station 2BN,Pre-Gondola II SD-l Event. 1610 Comb filter peak velocity as a function of frequency for Station 2BN,Pre-Gondola II SD-2 Event. 1611 Comparison of comb filter peak velocities of radial motion atStation 2BN for SD-l and SD-2 Events. 1712 Comparison of comb filter peak velocities of transverse motion atStation 2BN for SD-l and SD-2 Events. 1713 Comparison of comb filter peak velocities of vertical motion atStation 2BN for SD-l and SD-2 Events.14 Comb filter peak velocity as a function of frequency for Station 6W,Pre-Gondola II SD-l Event.15 Comb filter peak velocity as a function of frequency for Station 6W,Pre-Gondola II SD-2 Event.16 Comparison of comb filter peak velocities of radial motion atStation 6W, for SD-l and SD-2 Events.17 Comparison of comb filter peak velocities of transverse motion atStation 6W for SD-l and SD-2 Events.18 Comparison of comb filter peak velocities of vertical motion atStation 6W for SD-l and SD-2 Events.19 Comb filter peak velocity as a function of frequency for Station 2BN,Pre-Gondola II140-ton row charge.20 Comb filter peak velocity as a function of frequency for Station 3N,Pre-Gondola II140-ton row charge.21 Comb filter peak velocity as a function of frequency for Station 4N,Pre-Gondola II 140-ton row charge.22 Comb filter peak velocity as a function of frequency for Station 5N,Pre-Gondola II140-ton row charge.

18181919202021212222

-vi-


7/77

Figures (Continued)23 Comb filter peak velocity of radial components for all stations,Pre-Gondola II 140-ton row charge. 2324 Comb filter peak velocities for all stations of transverse components,Pre-Gondola II 140-ton row charge. 2425 Comb filter peak velocity of vertical components for all stations,Pre-Gondola II 140-ton row charge, as a function of frequency.(Note that at low frequencies attenuation is about as expected; athigh frequencies the short duration-high amplitude pulses insignals on Station 5N considerably alter the relationship ofV vs distance.) 25p26 Plot of a(frequency) from equation: CFPV(R, f) = KfR-a(f); thisrepresents attenuation slope for comb filter peak velocity as afunction of frequency (Pre -Gondola II). 2527 Radial components at Station 2BN for three yields. 2728 Transverse components at Station 2BN for three yields. 2829 Vertical components at Station 2BN for three yields. 2930 Resultant power spectral densities at Station 2BN for threeyields. 3031 Radial components at Station 6Wfor two source configurations. 3132 Transverse components at Station 6W for two sourceconfigurations. 3233 Vertical components at Station 6W for two sourceconfigurations. 3334 Resultant power spectral densities at Station 6Wfor twosource configurations. 3435 Radial components at Station 2BNfor two source configurations. 3536 Transverse components at Station 2BN for two sourceconfigurations. 3637 Vertical components at Station 2BN for two sourceconfigurations. 3738 Resultant power spectral densities at Station 2BN for two sourceconfigurations. 3839 Radial components for Pre-Gondola I Bravo Event at variousdistances. 3940 Transverse components for Pre-Gondola I Bravo Event atvarious distances. 4041 Vertical components for Pre-Gondola I Bravo Event at variousdistances. 4142 Resultant power spectral densities for Pre-Gondola I Bravo

Event at various distances. 4243 Radial components for Pre-Gondola II 140-ton row charge atvarious distances. 4344 Transverse components for Pre-Gondola II 140-ton row chargeat various distances. 4445 Vertical components for Pre-Gondola II 140-ton row charge atvarious distances. 4546 Resultant power spectral densities for Pre-Gondola II 140-tonrow charge at various distances. 46

-vii-


8/77

Figures (Continued)47 Radial components at Station 2BN for four 20-tonevents at different depths of burst. .48 Transverse components at Station 2BN for four 20-tonevents at different depths of burst.49 Vertical components at Station 2BN for four 20-ton events

at different depths of burst.50 Resultant power spectral densities at Station 2BN for four20-ton events at different depths of burst.51 Peak power spectral densities plotted against radius forPre-Gondola IBravo Event (a comes from the equation:PSDp(R) = KR-a).52 Peak power spectral densities plotted against radius forPre-Gondola II 140-ton row charge (a comes from theequation: PSDp(R) = KR-a).53 Energy in seismic signal for components and total energyas a function of radius for Pre-Gondola IBravo Event(a comes from the equation: :L:E(R) = KR-a).54 Energy in seismic signal for components and total energyas a function of radius for Pre-Gondola II 140-ton row charge(a comes from the equation: :L:E(R) = KR-a). .. . .55 Slope of power spectral density attenuation as a function offrequency for Pre-Gondola IBravo Event (i, e., slope acomes from the equation: PSD(R, f) = KfR-a(f. 56 Slope of power spectral density attenuation as a function offrequency for Pre-Gondola II 140-ton row charge (i, e.,slope a comes from the equation: PSD(R, f) = KfR-a(f.57 Comparison of attenuation for resultant power spectraldensities plotted as a function of frequency for Pre-Gondola I and II.58 Comparison of slopes of attenuation of comb filter peakvelocities to one-half of slopes of attenuation of powerspectral density resultant for Pre-Gondola II 140-ton rowcharge as a function of frequency.59 Total station energy (:L:E) at Station 2BN (Pre-Gondola I)showing transmitted energy dependence upon actual depth ofburst (Bravo datum is neglected-see text); curve isextrapolated to DOB's for 40-ton and 0.5-ton events toobtain comparisons for Pre -Gondola II.60 Total station energy (:L:E) vs yield after depth of burst effecthas been removed.61a Radial and transverse components at Station 2BN for Pre-Gondola I(all events).61b Vertical component at Station 2BNfor Pre-Gondola I(allevents).61c Resultant velocity at Station 2BNfor Pre-Gondola I(all events).62 Signals recorded at Station 2BN for three different yields(0.5- and 20-ton Single bursts and 140-ton row charge).

4748

4950

51

51

51

51

52

52

52

53

535354545555

-viii -


9/77

INTERMEDIATE RANGE GROUND MOTIONSFOR PRE-GONDOLA II ANDASSOCIATED EVENTSIntroduction

DESCRIPTION OF PROJECT PRE-GONDOLAIIProject Pre-Gondola II was a row-

charge cratering experiment in weak, wetclay-shale conducted by the U. S. ArmyEngineer Nuclear Crate ring Group (NCG)as a part of the joint Atomic EnergyCommission-Corps of Engineers nuclearexcavation research program. The pri-mary purpose of this nominal 140-ton. five-charge experiment was to gain row-charge cratering experience in a weak,wet medium. In addition, this experimenttested techniques for connecting a row-charge crater to an existing crater andfor over-excavating to accept throwoutfrom a follow-on connecting row-chargecrater.

Project Pre -Gondola II was detonatedat Valley County, near the edge of theFort Peck Reservoir approximately 18 misouth of Glasgow, Montana, at0800 hr (MDT), 28 June 1967 (see Fig. 1).Coordinates of the center charge wereW10638'31", N4755'51". The orientationof the row was along an alignment 10East of North to 10 West of South. Thisalignment extended through the center ofthe Charlie crater (an existing cratercreated during the Pre-Gondola I series).The average lip crest-to-lip crest

dimensions after connection to the Charliecrater were 640 ft X 280 ft. Individualcharge yields, depths, spacing, and re-sulting apparent crater dimensions aregiven in Table 1.

Table I, Pre-Gondola II charge yields, depths, spacing, and resulting apparent craterdimensions.Emplacement Apparentonfi,ura tionft) crater dimensionsDepth of (ft)Charge Tons NMa burst Spacing Width Wa Depth Da

Charlie (existing crater)b (19.62) (42.5) - (160.8) (32.6)E (connecting charge) 38.61 59.7 105.5c 206.5 55.5F 19.70 49.4 79.8 152.5 37.5G (center charge) 19.55 48.8 79.9 164.0 36.9H (over-excavating charge) 39.56 59.9 79.9 214.5 57.0I 20.00 48.8 79.9 173.0 33.5aSpherical charges of liquid explosive nitromethane (CH3N02). Actual total chargeweight was 137.42 tons.bDimensions for Charlie crater are those existing prior to row-charge event.cDistance between this charge and charge on line above.

-1-


10/77

~ Airport Paved Road... Landing Strip - - -- Graded Road

Scale in Mileso 5NI

Fig. 1. Map showing environs of Pre-Gondola II and associatedevents.

-2 -


11/77

DESCRIPTION OF EVENTS ASSOCIATEDWITH PRE -GONDOLAIIPrior to the row-charge detonation,

two 1000-lb shots were fired to investigatethe effects of small-magnitude decouplingfactors. These events were SD-1 (fully

coupled) and SD-2 (decoupled, V it /cavi yV = 2.0). These events were bothexpdetonated on 14 June 1968 at the same siteas the 140-ton event. Cratering data forthese two events are summarized asfollows:

Apparent crater dimensionsCharge Cavity volume DOB {ft)Event (tons NM) explosive volume (ft) Width Wa Depth Da---SD-1 0.50 1.0 17.3 50.2 10.5SD-2 0.50 2.0 17.7 47.4 10.5

DESCRIPTION OF OTHER EVENTS RE-LATED TO PRE -GONDOLAIIIn addition to the above events, con-

siderable additional analysis has beenmade of seismic records taken duringPre-Gondola I, located at the same siteas the Pre-Gondola II events. Thisanalysis is reported here. The work was

Charge DOBEvent (tons NM) (ft)Charlie 19.62 42.49Bravo 19.36 46.25Alfa 20.35 52.71Delta 20.24 56.87

OBJECTIVES

In order to provide a knowledgeablebasis to plan for possible large scale geo-nuclear projects in the future, it is nec-essary to gather information on thecharacteristic behavior of various earthmaterials at the shot point and in thevicinity of subsurface explosions. ThePre-Gondola series of explosions offersan opportunity to study the effect of severalimportant parameters on the transmitted

originally reported in PNE-ll05.1 Sincethe writing of that report additionalanalysis techniques have become available,and it is imperative that for a completeunderstanding of the results of Pre-Gondola II, these data and results be in-cluded at this time. A summary of thePre-Gondola I cratering data is as follows:

Apparent crater dimensions(ft)Width Wa Depth Da160.8 32.6157.0 29.5152.2 32.1130.2 25.2

seismic signals. The varietyof yields, depths of burst, and shot config-urations, all being detonated atessentially the same geographical location,makes it possible to make repeated mea-surements at the same locations. Thisprocess means that the effect of trans-mission path can, to a large extent, benormalized out of the data at any givenstation. (This is not entirely true sincethe "transmissibility" of the transmissionpath is frequency-dependent and, therefore,

- 3-


12/77

may "change" if the source function isaltered drastically with respect tofrequency-see Appendix A.)The explosions were detonated in a wet

clay-shale media which is characteristicof several locations of interest to the nu-clear crate ring program. This region ofnortheastern Montana is also of interestbecause it appears to be relatively simplefrom a physical properties standpoint.Flat-lying, extensive, thick layers ofsedimentary materials exist to depths ofover 8,000 ft and distances of over 100 km.In view of the above, the following objectiveswere chosen for this study:l. Dependence of velocity time historyof ground motion upon yield, depth ofburst, and various shot configurations(single charge, row-charge, and smalldecoupling ratios).2. Dependence of peak velocity spectra

and power spectral densities upon yield,depth of burst, and shot configuration.3. Identification of those properties of

explosively induced seismic signals whichscale uniformly and consistently over largeranges of yield, depth of burst, and varia-tions in shot configurations.4. Empirical measurements of seismic

motion from cratering explosions in an ex-tensive water-saturated clay-shale mate-rial.5. A study of the attenuation with dis-

tance of peak signal amplitudes, velocityspectra, and power spectral density in asimple geology.

BACKGROUND

Aword of explanation is necessary atthis point. Although this report is tech-nicallya report on Pre-Gondola II, it also

contains much information gathered duringPre -Gondola 1. The reasons for this aretwo-fold: (1) the events from the separateprograms are meaningful to the fullestextent only when correlated and presentedas a unit, and (2) the power spectraldensity analysis technique (PSD) was notavailable when the Pre-Gondola I work wasreported. Itwas the conclusion of thatreport (PNE-ll05) that, when the analysistechnique did become available, the datashould be reported. Itis included in thisreport to simplify the task of both thereader and writer.Figure 1 shows the location of the Pre-

Gondola experiments in northeasternMontana. The ideal geology for seismicpropagation studies would be a simple two-layer medium, flat, and of infinite extentin the horizontal directions. This is thesimplest geometry that would allow forpropagation of surface waves, direct p ands arrivals and refracted arrivals. Theauthor is unaware of the existence any-where of such a geologic setting ofappropriate dimensions.The geology of northeastern Montana is

however relatively simple.2 Massivesedimentary beds make up the surfacegeology. These vary in thickness up to2000 ft and comprise in total thicknessabout 5000-8000 ft. Structural featuressuch as the Williston Basin are of the orderof hundreds of miles across and of relative-ly shallow relief, resulting in relativelyflat bedding planes.Previous work on seismic attenuation

has been primarily concerned with peakamplitudes of motion in the wave trainswith little or no regard for the total fre-quency content of the motion.3,4 The rea-sons for this condition have been twofold.

....-


13/77

Until recently there has been a lack ofgood instrumentation and analysis capa-bilities to do this work. More important,most of the experience to date has been insuch complicated geologic settings thatmeaningful frequency analysis has beenalmost impossible. The relatively simplegeology offers an opportunity to investigateboth the frequency content and the amplitudeas a function of distance from the sourceand for variations in yield and depth ofburst of the source.Most explosions with yields greater

than 5 tons have fallen into three categories:1. Industrial explosions of all kindswith usually little or no instrumenta-tion.

2. Well-instrumented, completely con-tained single-everit explosions.

3. Well-instrumented cratering exper-iments.

For the most part the larger crateringexperiments prior to Pre-Gondola havebeen single events. Where there havebeen several events, the layout andseismic instrumentation offered noopportuni ty to study coupling of ener gyinto the seismic signal.A requirement for a dynamic cratering

condition (as opposed to a subsidencecrater *) is that the cavity communicateswith the atmosphere before the cavitygrowth has stopped. This means that the

overburden must fail in tension, and thethus-isolated blocks of overburden will beejected along a ballistic trajectory.Seismically, this differs from the com-pletely contained situation by allowing thecavity pressure to drop at early times,thus preventing the coupling of somefraction of the energy into the seismicSignal at late times. The possible de-crease in seismic energy is an unknown,and previous estimates of seismic motionsfor cratering shots have been based on re-sults for the completely contained case.

If there is an effect due to cratering itcould be proportional to a function ofeither the actual depth or the scaled depth(D/w1/3.4). In either case some effectwould be noticed for a shallow, low-yieldexplosion. If coupling is a function ofactual depth (in other words, a function ofthe total time before breakout), one wouldexpect to see a less marked effect fordeeper. high-yield events. If the effectis a function of the scaled depth, then itshould be the same for all yields. Pre-Gondola II and associated events offer asuitable variety of events to study thesepossibilities.

* .subaidence crater occurs when anexplosively created subsurface cavitycollapses resulting in a surface depres-sion without any direct cavity-atmosphereinteraction.

-5-


14/77

Procedure

STATION SELECTION

The number and location of seismicstations for Pre-Gondola II were essentiallythe same as for Pre -Gondola I, since thecriteria were identical. 1 Figure 2 is amap showing the location of stations forthe events reported herein. All stations,their locations with respect to SGZ, andthe events during which they were occupiedare given in Table II along with other infor-mation. This coverage is sufficient toprovide the following:

1. Five stations ona radius for twodifferent yields (one of them a rowcharge)

2. One station repeated for a singleyield at four depths of burst

3. Two stations repeated for the de-coupling sequence

4. One station repeated for three yieldsat optimum depth of burst

5. Stations 90 deg apart with respectto SGZ to study the asymmetry ofseismic radiation from a rowcharge.

Table II. Events reported and data analyses used for velocity geophone measurements.Station location Quality Analysis tech-Event Station Distance Azimuth of data niques use d=

Pre -Gondola IAlfa 2BN 11.65 8E of N Excellent VA, PSDCharlie 2BN 11.65 8E of N Excellent VA, PSDDelta 2BN 11.65 8E of N Excellent VA, PSDBravo 2BN 11.65 8E of N Excellent VA, PSDBravo 3N 29.3 9E of N Poor VA, PSDBravo 4N 52.0 8E of N Good VA, PSDBravo 5N 95.9 6E of N Excellent VA, PSD

Pre -Gondola IISD-1 2BN 11.65 8E of N Fair VA, CFPV, PSDSD-1 6W 6.39 900W of N Good VA, CFPV, PSDSD-2 2BN 11.65 8E of N Fair VA,CFPV,PSDSD-2 6W 6.39 900W of N Good VA,CFPV,PSDRow charge 2BN 11.65 8E of N Excellent VA, CFPV, PSD(E,F,G,H,!) 3N 29.3 9E of N Excellent VA, CFPV, PSD

4N 52.0 8E of N Excellent VA,CFPV,PSD5N 95.9 6E of N Excellent VA,CFPV,PSDIN -5.0 8E of N Very poor None6W 6.39 900W of N Very poor None

aVA = vector addition, PSD = power spectral density, CFPV = comb filter peakveloci.ty.

-6-


15/77

10Miles

I 20 Kilometers Ground Zero Seismic Stations_ Paved Roads___ Dirt Roads

Fig. 2. Map showing location of free-field seismic stations for Pre-GondolaII 140-ton row charge. (Station 2BNwas operated for Alfa, Charlie,Delta, and Bravo Events, and Stations 3N, 4N, and 5Nwere occupiedon Bravo Event during Pre-Gondola series; Stations 2BNand 6Wwereoccupied on decoupling series during Pre-Gondola II.)

-7 -


16/77

SEISMICINSTRUMENTATION

Ground motions were measured withmodified HS-IO-I velocity geophones. TheHS-IO-I is a moving coil instrument witha I-sec natural period; with 700/0dampingit has a flat response curve above I Hz(cycles per second). The sensitivity isdetermined by the coil impedance (Fig. 3).To improve performance and reliabilityand to reduce instrument drift, adjustmentswere made in the field and all instrumentswere monitored after emplacement.Emplacement was accomplished aboutI wk prior to the event date, and the finalmonitor check was made on the eveningprior to the event date. The critical

UQIIII>'-...c

-JQ.-Jo

parameters are sensitivity, natural period,and damping ratios.Each station consisted of one vertical

and two horizontal instruments, the latteroriented radially and transversely to theshot point. This comprises the standardrectangular system, giving motion in thethree mutually orthogonal directionsnecessary to define translational motion.As a matter of definition, for the rest ofthis report the following designations willbe used: x refers to radial motion, yrefers to transverse motion, z refers tovertical motion, and Res is the vector sum

( J x2 + y2 + z2) for velocity and anarithmetic sum (x + y + z) for energy.Isolation from wind noise and good

mechanical coupling with the earth are

Model - 123K verticalNatura I frequency - 1.04 HzCoil resistance - 125,000 rlOpen circuit damping - 35%

Open = 35%damping

0.66 M rl= 70%dam in

Frequency - H zFig. 3. Response of HS-IO-I velocity geophone at 35 and 700/0 damping.

-8-


17/77

accomplished by burying the instruments.To meet the requirements of instrumentprotection, ease of operation and emplace-ment, quality of orientation, etc., aninstrument canister was developed tomaintain the instruments in the mutuallyorthogonal relationship. The water-tightcanister is then easily buried in a 12-in.diameter, 40-in. hole and oriented as aunit. Tamped sand was used as a backfill.Figure 4 illustrates the details of theinstrument package.

RECORDINGANDTIMING

Magnetic tape recorders utilizing anFM electronics system, a four-track head,

and a 1/4-in. loop-tape cartridge with a5-1/2-min capacity were used on allmeasurements. One of these recordersis shown in Fig. 5 and partly dismantledin Fig. 6. The recorder shown was usedon the Pre-Gondola I Events Alfa, Bravo,Charlie, and Delta. For the Pre-GondolaII Events certain modifications were madeon the power supplies and VCOIS to im-prove their temperature and timestability. These modifications did notsignificantly alter the appearance of therecorders.Each geophone (component) was recorded

on one track of the magnetic tape and anoscillator-controlled timing signal wasrecorded on the fourth track. An

Fig. 4. Three-component instrument package during set-up and prior to in-sertion into outer canister (canister is partly visible behind metaldrum on right).-9-


18/77

Fig. 5. Top view of tape recorder used for recording geophone signals (tape cartridgeis shown in place at right).

approximate zero time signal (squarepulse) was manually superimposed on thetiming channel by the station operator whowas in radio contact with the firing controlcenter.The cross timing accuracy between

stations is determined by the variation instation operator reaction time. Thisvariation could cause an uncertainty of asmuch as 1 sec in actual zero time. Thisuncertainty is considered too large toallow meaningful analysis of the transittimes of various wave phases, etc.

INFORMATIONRETRIEVAL ANDREDUCTIONInstrument responses to step functions

were recorded prior to events. Recorder

calibration signals (sine waves and ramps)were recorded on the "event tapes" beforeand after the event. All recording equip-ment was turned on and running at minus30 sec. A positive de voltage of aboutI-sec duration was manually superim-posed at zero time on the timing channel.

Subsequent to the event, magnetic taperecords from all stations were redubbedto a I-in. tape to make them compatiblewith the digital analysis system. Recordswere digitized with an Astro-Data analog-to-digital translator. Analysis ofdigitized records was completed onIBM7094 and CDC 3600 computers. Thedigitized record contained 400 data pointsper sec for each component.

-10-


19/77

DATAANALYSIS

The most used method of analyzingseismic motion data of the type withwhich we are concerned has, in the past,been to measure the peak amplitudeattained regardless of where in the recordit occurred. Such an analysis is crude tobe sure, but little else could be done dueto the high cost of hand analysis on themany records that had to be studied.Digital computers have become commonenough to reduce the total cost and more

sophisticated analysis techniques are nowpossible.

It is of interest at this time to comparethis method to newer ones and to obtainsome measure of the validity of the oldmethod. Accordingly peak motions weremeasured from the records. Also mea-sured were the time of arrival of thatmotion after the first motion and theapproximate frequency of the motion.The latter is calculated by assuming thehalf wave to be of harmonic nature; i.e.,

Fig. 6. Partly dismantled recorder, showing its size, portability, etc.

-11-


20/77

f = (1/2) (1/~t) where ~t is the time be-tween zero crossings just prior to andfollowing the peak motion.One of the newer techniques for

analyzing a time varying signal is to filterit into its spectral components to mea-sure the peak amplitudes in each compo-nent generating a velocity spectra. Thismethod is called comb filter peakvelocity (CFPV) and is described inAppendix B.The actual equipment used in this

analysis was band pass filters. Thesignal ratios at 1 octave from the centerfrequency were about 15 dB for power(voltage. /voltage t::::6). Center fre-in ouquencies were set in a semi -systematicpattern to cover the range of interest.They were 0.75, 1.0, 1.5, 2.0, 3.0, 4.0,5.0, 6.0, 8.0, 10.0, 12.0, and 15.0 Hz,respectively.

Comb filtering was planned for therecords obtained in the 140-ton rowcharge, as well as the SD-1 and SD-2Events. The peak amplitudes for eachfrequency from a given geophone wereplotted against frequency to illustrate thespectral content of the signal. Comparisonof the plots of the three space coordinates,the different stations, and the severalevents provides an understanding of howthe spectral content varies with distance,space direction, source configuration (forthe SD-1 Event compared to the SD-2) andfor yield (the SD-1 Event comparee to the140-ton row charge). Attenuation forvarious frequencies can be determinedfrom a fit to the data. The equation is

Af = Amplitude (f) = Kf R-a(f)where a is the attenuation slope.A comparison of these results to those

obtained by the power spectral densitymethod of analysis (PSD) provides a basisfor choosing the most consistent andappropriate method of the two for analyz-ing ground motions from undergroundexplosions.The three rectangular components are

sufficient to describe the translationalmotion of a point. However to gain aknowledge of the absolute amplitude ofmotion requires a transformation tospherical coordinates. The theory andcomputer code for the transformation isdescribed in Appendix B. 2The code calculates x, y, z, (Res) ,IRes L e , and < j> as functions of time. Itis called (surprisingly) VECTOR.Another method of analyzing a signal

to obtain a measure of its spectral con-tent and total energy is to determine thePSD and the integral of (PSD)df or E.This method is described in Appendix Cand Ref. 5. The quantities produced bythe PSD code for a given time history ofvelocity are PSD(f), or density of poweras a function of frequency for threecomponents, a resultant E, or totalenergy in a component, and LE or totalstation energy.The PSD's for all signals in this report

were to be calculated and comparisonsmade to determine the variance with yield,depth of burst, source configuration, andradius and to make selected comparisonswith the CFPV's for some events andstations.

-12-


21/77

ResultsDATARECOVERY ANDREDUCTION

Table II lists the events included inthis report along with the various seismicstations, their locations with respect toSGZ, the quality of the data recorded,and the analysis techniques used on eachstation recording.The failure of the primary recorder

is the reason for the poor quality onStation 3N for Bravo. A backup recorder.set at low gain recorded the event but thesignal-to-electronic noise ratio is small.On the 140-ton row charge the motionsat Stations 6W and 1N exceeded themechanical limits of the geophones,making the data totally unreliable. Noresults for these two stations on thisevent will be included. The small motionsfrom Events SD-1 and SD-2 resulted in asmall signal-to-background seismicnoise, reducing the quality of these signalssomewhat.The method of data reduction requires

redubbing which introduces additionalnoise, usually random or white, butsometimes peculiar to a given machineas a result of mechanical flutter in thetape drive, etc. This is apparently thecause of spikes in the power spectraldensity curves at 9-1/2 and 15Hz in all ofthe SD-1, SD-2, and 140-ton row-chargeEvents from Pre-Gondola II. The largespike in the PSD for all components ofStation 3N on Bravo is not present for the140-ton Event at 3Nnor is it present onany other station. Itis assumed that thisis due to the tape drive of the event re-corder especially since this was the lowgain recorder. There is, however, some

doubt about the validity of this assumptionand both possibilities are reported in thepeak amplitude values. It has little effectupon the total energy E (see Appendix C)in the signal.

PEAK VELOCITY ANDVECTORADDITION

Table III lists the peak velocity of thecomponents and the peak vector resultant,the time of arrival of that peak after firstmotion, and the apparent frequency ofthat motion for all stations and all events.Peak velocities for the stations recordingfor Bravo and the 140-ton row charge areplotted in Figs. 7 and 8 as a function ofdistance from SGZ.

COMB FILTER PEAK VELOCITY

Only the records from SD-1, SD-2,and the 140-ton row charge were analyzedin this manner. Table IV contains thesedata. A better concept of these data isgained when they are put in graphicalform. Figures 9 and 10 show the resultsfor Station 2BNon SD-1 and SD-2, re-spectively, and Figs. 11, 12, and 13 com-pare the three separate space coordinatesfor the two events to illustrate the differ-ences due to the small decoupling factor.Figures 14 through 18 show results forStation 6W for the same events in thesame sequence.The CFPV -frequency curves for each

station for the 140-ton row-charge areshown in Figs. 19 through 22, while theindividual components for all stations arecompared in Figs. 23 through 25. The

:..13-


22/77

,Table III. Velocity data-all stations.Time of arrival Frequency

Peak velocity after first motion of motionEvent Station Component (mm/sec) (sec) . (Hz)SD-1 2BN x 0.0385 11.2 1.75SD-1 2BN Y -0.027 11.6 1.75SD-1 2BN z 0.0338 0.8 6.8SD-1 2BN Res 0.042 11.6SD-2 6W x 0.108 7.6 2.4SD-2 6W Y 0.096 9.4 3.7SD-2 6W z 0.096 9.97 8.0SD-2 6W Re~ 0.113 7.6SD-2 2BN x 0.0545 11.9 1.75SD-2 2BN Y -0.045 12.5 1.75SD-2 2BN z 0.0432 1.4 7.6SD-2 2BN Res 0.0687 12.6SD-1 6W x 0.055 7.5 2.5SD-1 6W Y 0.072 7.5 3.4SD-1 6W z 0.099 1.0 8.8SD-1 6W Res 0.099 1.0Bravo 2BN x 0.76 12.4 1.75Bravo 2BN Y 0.34 12.4 1.2Bravo 2BN z 1.20 2.4 4.5Bravo 2BN Res 1.29 2.4Bravo 3N x 0.16 ? 1.5Bravo 3N Y 0.10 ? 1.5Bravo 3N z 0.125 ? 2.5Bravo 3N Res 0.220 ?Bravo 4N x 0.0207 12.65 3.0Bravo 4N Y 0.0142 14.7 3.0Bravo 4N z 0.0257 13.5 2.5Bravo 4N Res 0.0306 13.5Bravo 5N x 0.024 10% 5.48 7.0Bravo 5N Y 0.014 10 " /0 5.53 8.0Bravo 5N z 0.08 10% 4.2 8.0Bravo 5N Res 0.08 10% 4.2Alia 2BN x 1.10 2.32 6.0Alia 2BN Y 0.45 11.9 2.0Alfa 2BN z 1.94 2.25 5.0Alia 2BN Res 2.09 2.25Charlie 2BN x 0.75 3.2, 11.2 5.0, 1.5Charlie 2BN Y 0.28 12.8 1.5Charlie 2BN z 1.38 3.1 5.0Charlie 2BN Res 1.46 3.1Delta 2BN x 1.72 11.8 1.75Delta 2BN Y 0.72 13.5 1.75Delta 2BN z 2.20 2.26 6.00Delta 2BN Res 2.45 2.34Row charge 2BN x 4.4 10.7 2.0Row char-ge 2BN y 2.8 11.7 0.8Row charge 2BN z 5.8 2.21 4.8Row charge 2BN Res 6.3 2.21Row charge 3N x 0.251 10.5 2.0Row charge 3N Y 0.209 20.8 1.75Row charge 3N z 0.655 6.8 3.0Row charge 3N Res 0.663 6.8Row charge 4N x 0.193 12.6 3.1Row charge 4N y 0.14 14.2 3.0Row charge 4N z 0.32 12.0 2.7Row charge 4N Res 0.344 12.2Row charge 5N x 0.097 5.4 4.0Row charge 5N y 0.046 12.9 3.3Row charge 5N z 0.197 4.0 7.0Row charge 5N Res 0.197 4.0

-14-


23/77

100~--------------------~

oQ)~EEI~. -s~ 0 . 1..)/.c~

1 0

A Radical (~) componentD Transverse ( y ) componento Vertical (z) component Resultant

Least squares Ifit (linear) tothe peak resu tants

O . 0 1 L-- ..___ .l..-~-~1 1000

oQ)~EEI~.-~~ O . 1~c&

1 0

Aoo

Radial (x ) componentTransverse ( y ) componentVertical (z ) componentResultantLeast squares. fit (Iineor) toresu I tant peaks

1 0Radius - km

Fig. 7. Peak velocities for Pre -Gondola IBravo Event, as a function of dis-tance (large value at 96 km is dueto single pulse of very highamplitude in velocity-time signa-ture; rest of signature is of rela-tively low amplitude).Table IV. Comb filter peak velocity data for Pre -Gondola II.

o

0.01L------~~----~------~Radius

1 0 0km

1000

Fig. 8. Peak velocities for Pre-Gondola II140-ton row charge as a functionof distance.

Event Station0.19

Component 0.75 Hz 1.0Hz 1.5 Hz 2.0 Hz 3.0 Hz 4.0 Hz 5.0 Hz 6.0 Hz 8.0 Hz 10.0 Hz 12.0 Hz 15.0 Hz140-ton 2ENrow charge

3N

4N

5N

x 1.72 2.870.67 0.780.97 1.84

2.101.231.300.270.150.36

0.860.564.000.21 0.200.11 0.120.63 0.45

0.35 0.23

1.620.335.40

2.100.675.400.140.110.22

1.910.564.64

1.530.472.38

0.960.331.73

0.480.230.~70.0210.0120.045 0.033

0.120.54

0.021 0.0100.0090.076 0.0850.022

0.13 0.068

2.871.342.16

yzx 0.23 0.18

0.170.20

0.083 0.062 0.0410.110.18

0.047 0.0240.13 0.078

SD-l 2EN

0.25y 0.094 0.19

0.19 0.310.052 0.042 0.057 0.12 0.18 0.15 0.083 0.052 0.037 0.0260.029 0.046 0.032 0.072 0.13 0.075 0.046 0.032 0.023 0.013

0.065 0.051 0.070

zxyz 0.065 0.054 0.081 0.20

0.022 0.022 0.039 0.044 0.039 0.033 0.033 0.0280.013 0.027 0.013 0.040 0.034 0.013

0.13

x 0.022 0.0110.013 0.007

0.011 0.045 0.068 0.057 0.085 0.11yz

0.11

0.14

0.008 0.010 0.031 0.036 0.019 0.013 0.008 0.010 0.013 0.008y 0.005 0.005 0.014 0.014 0.011 0.007 0.005 0.011 0.009 0.005

6W x0.005 0.005 0.012 0.012 0.010 0.014 0.017 0.022 0.026 0.0190.012 0.018 0.039 0.058 0.066 0.066 0.048 0.051 0.042 0.0300.007 0.014 0.018 0.053 0.064 0.043 0.021 0.028 0.0180.008 0.011 0.039 0.044 0.044 0.048 0.033 0.048 0.069 0.066

z

yz

0.014 0.0090.027 0.0120.0140.066 0.061

SD-2 2EN

6W

y 0.0050.005

0.008 0.019 0.050 0.053 0.029 0.015 0.015 0.015 0.017 0.0090.0050.034

0.009 0.023 0.027 0.014 0.0140.012 0.026 0.026 0.014 0.017

x

z0.011 0.0070.024 0.034

0.0180.036

x0.007 0.018 0.036 0.064 0.078 0.0500.008 0.025 0.050 0.056 0.061 0.044

0.012 0.030 0.060 0.072 0.072 0.072 0.042 0.036 0.042 0.0330.0140.030 0.0150.0140.069 0.061

yz

-15-

0.0280.033

0.0250.050

0.025 0.0210.066 0.077


24/77

o 0Radial (x ),I!r- - - ~ Transverse (y )0-----0Vertical (z)

.4, -.s : t ' .p' q. quu"EI.~ 0.01M"'ii ., .>~0~

\o

o.OO1L---------------~--------------~--------------~1000.1 10Frequency - Hz

Fig. 9. Comb filter peak velocity as a function of frequency for Station 2BN, Pre-Gondola II SD-1 Event.

0.001~--------------~----------------~--------------~O.1 10 100

0.1r-----------------~------------------~----------------~o oRadial (x)/:r-----i1 Transverse (y )o-._._.oVertical (z)r'~i \. \

~

uUtil .. 0.01-u0"'ii>~

Frequency - HzFig. 10. Comb filter peak velocity as a function of frequency for Station 2BN,Pre-Gondola II SD-2 Event.

-16-


25/77

o.oo1L--------- -L ~--------------~O.1 10 100

0Q)III- 0.01-00Q)>..:,t.cQ)0..

o 0 SD-l0-----0 SD-2

Frequency - HzFig. 11. Comparison of comb filter peak velocities of radial motion at Station 2BNfor SD-l and SD-2 Events.

0 . 1 r - - - - - - - - - - - - - - - - - - - - - - - , ~ - - - - - - - - - - - - - - - - - - - - - _ . - - - - - - - - - - - - - - - - - - - - - - - - _ ,

0.001L---------------------L--------------~1~0~------------------~100O . 1

0Q)..:,t.cc f

0----0 SD-l0----0 SD-2

Fig. 12. Comparison of comb filter peak velocities of transverse motion at Station2BNfor SD-l and SD-2 Events.

Frequency - Hz

-17-


26/77

o.1r-----------------.---------------~-----------------,0--0 SD-1(>.---0 SD-2

u~


27/77

0 . 1 r - - - - - - - - - - - - - - - - - r - - - - - - - - - - - - - ~ ~ - - ~ - - - - ~ ~ ~ - - ~

UG JII>- 0.01. . . .'u~G J>..::t.C& .

0.001~--------------~----------------~--------------~0.1 10 100Frequency - Hz

Fig. 15. Comb filter peak velocity as a function of frequency for Station 6W,Pre-Gondola II SD-2 Event.

0.1.----------------.----------------~--------------~.o--~"p -IIIP

o--oSD-l0-----0SD-2

UG JII>- 0.01. . . .'u.sG J>~CG Ja..

0.0011O . 10 100Frequency - Hz

Fig. 16. Comparison of comb filter peak velocities of radial motion at Station 6W,for SD-1 and SD-2 Events.-19-


28/77

O.1. . . . ~P ' 0 050-1I 0-----0 50-2IPo IQ) III> I ~f ~I I

>. . 0.01 I- I0. . .2 &)> \2Q..0.001~--------------~----------------~ ~ ~O . 1 10 100Freouencv - HzFig. 17. Comparison of comb filter peak velocities of transverse motion atStation 6Wfor SD-1 and SD-2 Events.

0.1r-----------------r----------------,----------~-- __~

oQ)."~e&

0---050-10-----050-2

0.001~---------------L----------------~---- ~O.i 10 100Frequency - Hz

Fig. 18. Comparisonof comb filter peak velocities of vertical motion at Station6Wfor SD-1 and SD-2 Events.

-20~


29/77

o.1L-----------------~--------------~10----------------~100O . 1

0.01~----------------~----------------~-- ~O.1 Frequency _ Hz 10 100

(JQ)'"- O . 1. - : :(J. . . 2Q)>~0Q)0..

10r-----------------r---------------~--------~------_,

(JQ)'"-+-(JoQ)>~oQ)0..

Frequency - HzFig. 19. Comb filter peak velocity as a function of frequency for Station 2BN,Pre-Gondola II 140-ton row charge.

A.I ~P

o 0Radial (x)b-----I:>. Transverse (y)Q-_-.oVertical (z)

Fig. 20. Comb filter peak velocity as a function of frequency for Station 3N,Pre-Gondola II 140-ton row charge.

-21-


30/77

0----0/!r-----Il0---0Radial (x)Transverse (y)Vertical (z)A.

I \

0.1u!


31/77

uQ)en-+-' 0oQ)>

. . : : LgQ..

10r-------------------r-------------------r-------------~--~

O . 1

5N

2 BN

3N

4N

0.01L- L_ L _ ~ - - - - - - - - - - - - - - ~O.1 100

Frequency - Hz

Fig. 23. Comb filter peak velocity of radial components for all stations, Pre-Gondola II140-ton row charge.

attenuation Of, (see Appendix B) withdistance for the peak amplitudes from theCFPV's is shown in Fig. 26.

total energies (E andLE.) for each com-ponent on all stations and events.

POWER SPECTRAL DENSITYComparisons of the PSD's for each

component (including the resultant) atStation 2BN for three yields are shown inFigs. 27 through 30. Comparisons of thePSD's for each component and resultant

Table V lists the peak density (PSDmax>'the frequency at which it occurs, and the

-23-


32/77

10~------------------~------------------~------------~--~~

0.01

uQ)"EI.~ 0.1s~

..:. i:o&

2B N

3N

4N

5N

?

o.oo1L-------------------L-----------------~1~0------------------~100o . 1Frequency - Hz

Fig. 24. Comb filter peak velocities for all stations of transverse components,Pre-Gondola II 140-ton row charge.

-24-


33/77

10.-----------------.-----------------.----------------.

u. .- - - :EEI

Fig. 25. Comb filter peak velocity of vertical components for all stations, Pre-Gondola II140-ton row charge, as a function of frequency. (Note that at low frequenciesattenuation is about as expected; at high frequencies the short duration-highamplitude pulses in signals on Station 5N considerably alter relationship of Vvs distance.) p

2BN

4 . 0 ~ - - - - - - - - - - - - - - - - - - ~ - - - - - - - - - - - - - - - - - - . - - - - - - - - - - - - - - - - - - - ,

1 :-u.. 2. .>.s:e. ,c,

3N

1.0~------------------~--------------------~--------------------~o . 1 10 100Fig. 26.

0.14N

Frequency - Hz

R

z (Vertical) 0----0x (Radial ) x----~y (Transverse) 0-__0\

d

Frequency - HzPlot of a (frequency) from equation: CFPV(R, f) = K R-a(f); this representsattenuation slope for comb filter peak velocity as a function of frequency(Pre -Gondola II).

-25-


34/77

Table V. Power spectral density: peak density values and total energy.Peak densit:l:': Total energyValue Fr-equency'" (E and L : E ~Event Station Component (mm/sec)2 sec (Hz) (mm/sec)

Pre-Gondola II 2BN x 0.805 1.17 0.633 E140-ton row y 0.164 1.50 0.250 Echarge z 0.255 1.33 0.570 ERes 1.20 1.17 1.453 ~E

3N x 2.26 X 10-3 1.87 5.65 X 10-3Y 2.72 X 10-3 1.67 2.89 X 10-3z 7.74 X 10-3 2.67 8.06 X 10-3Res 10.5 X 10-3 2.67 16.60 X 10-3


5N x 1.54 X 10-4 2.83 2.83 X 10-4y 0.274 X 10-4 4.67 1.36 X 10-4z 2.57 X 10-4 2.67 3.51 X 10-4Res 4.30 X 10-4 2.67 7.70 X 10-4

SD-1 2BN x 4.68 X 10-5 1.83 4.4 X 10-5y 2.51 X 10-5 1.63 4.1 X 10-5z 1.62 X 10-5 1.50 4.1 X 10-5Res 7.42 X 10-5 1.83 12.6 X 10-5

SD-2 2BN x 11.40 X 10-5 1.83 8.8 X 10-5Y 5.12 X 10-5 1.67 7.1 X 10-5z 4.06 X 10-5 1.50 7.7 X 10-5Res 17.4 X 10-5 1.83 23.6 X 10-56W x 2.492 X 10-4 1.67 5.12 X 10-4

Y 3.28 X 10-4 2.17 4.24 X 10-4z 2.35 X 10-4 3.00 3.84 X 10-4Res 6.35 X 10-4 2.17 13.2 X 10-4

SD-1 6W x 0.64 X 10-4 2.17 1.31 X 10-4Y 1.12 X 10-4 3.67 1.95 X 10-4z 1.29 X 10-4 3.00 2.67 X 10-4Res 2.45 X 10-4 3.16 5.93 X 10-4

Pre-Gondola I 2BN x 2.17 X 10-2 1.17 2.40 X 10-2Bravo y 0.66 X 10-2 1.17 0.70 X 10-2z 0.97 X 10-2 1.33 2.05 X 10-2

Res 3.61 X 10-2 1.33 5.15 X 10-23N {9.09 X 10-4 7.50 29.03 X 1O-4bx 8.06 X 10-4 1.50{4.37 X 10-4 7.50 12.91 X 1O-4bY 2.17 X 10-4 1.83

z {3.94 X 10-4 7.50 14.69 X 1O-4b3.43 X 10-4 2.67Res {17.40 X 10-4 7.50 56.63 X 1O-4b9.57 X 10-4 1.50



Alfa 2BN x 7.36 X 10=~ 1.33 8.58 X 10-2Y 0.736 X 10_2 1.87 1.47 X 10-2z 2.46 X 10 1.50 6.91 X 10-2Res 10.26 X 10-2 1.33 16.96 X 10-2

Charlie 2BN x 2.45 X 10-2 1.33 2.67 X 10-2Y 0.535 X 10-2 1.67 0.65 X 10-2z 1.16 X 10-2 1.33 3.02 X 10-2Res 3.96 X 10-2 1.33 6.34 X 10-2

Delta 2BN x 9.53 X 10-2 1.33 10.46 X 10-2y 2.29 X 10-2 1.67 2.16 X 10-2z 3.36 X 10-2 1.33 7.68 X 10-2Res 13.60 X 10-2 1.33 20.30 X 10-2

aFrequency at which peak density occurs.bSpike at 7-1/2 Hz is on all three channels and unexplained at this time.

-26-


35/77

oQ)'" 10-3N_oQ).zEE>- . 10-4. . . . .'"cQ)"U0,_. . . . .oQ)

1 0 - 5.'"_Q)~0Q..

10-6

/140.ton~ / N~!\

!~,II I.: . !"'I"", I": 1~ 1. .~')\ : . . . . . L~ "" , -I \ _.\ , '' ' '' ' J l " I / \ ' . .- . . .. . , ., . . . . .. . . .


36/77

10-1

140-ton

10-2uQ)'"N-Q)

" 10-3EE> ..- DeltaIIcQ)""00 10-4L--UQ)a.IIIL-Q)

~0.. ." . . . . " .. .10-5

- . . . . . . . . . . . . . . . . .

1 0 r - - - - - - - - - - - - - - - - - ~ - - - - - - - - - - - - ~ ~ ~ - - - - - - - - - - - - - - ~

10-6~--------------_+--------~~~~~~------------~

10-7~ ~ ~U_ ~10-1

Frequency - Hz

Fig. 28. Transverse components at Station 2BNfor three yields.

-28-


37/77

10r--------------r-------------.------------~

10-1 /n1 4 0 - , o n )10-2 J f \\

uQ)'"N- - - -Elu 1 0 - 3~D e l I O !

~I> -:!:'"cQ)."C. . .. . . . 1 0 - 4Q)c ..II). . .Q)30c,

1 0 - 5 !\I ISD-1J \ A ~1 0 - 6 1 1V u I I

10-7~ ~ ~ ~10-1

Frequency - Hz

Fig. 29. Vertical components at Station 2BNfor three yields.

-29-


38/77

1 0 - 2oG)onN.,-...

EluE ~._..>. . 1 0 - 3. - : : DeltaoncG)"E. . .oG)a.on. . 1 0 - 4)~8 .. . .c0. . .:>onG)

0.::1 0 - 5 50-1

10-7~--------------~----------------~----------------~10-1

Frequency - Hz

Fig. 30. Resultant power spectral densities at Station 2BNfor three yields.

-30-


39/77

at Stations 6Wand 2BN for SD-1 and SD-2Events are shown in Figs. 31 through 38.Comparison of PSD's for each compo-

nent at various distances are shown inFigs. 39 through 42 for the Bravo Event,and in Figs. 43 through 46 for the 140-tonrow charge. Figures 47 through 50 showcomparisons of the PSD's of the separate

components at Station 2BNfor the 20-tonPre-Gondola Events Alfa, Bravo, Charlie,and Delta to show the effect of depth ofburst.The peak PSD's for Bravo and the 140-

ton row charge are plotted against distancein Figs. 51 and 52, respectively. Thetotal energies, E and :EE, are shown in

uQ)'"N_UQ)' " 1 0 - 5EE>- .. . . . .'"cQ)"'00 1 0 - 6. .. . . . .uQ)c. .'". .Q)~c . . .

1 0 - 7

- - S D - 2

/>:- - - S D - 1I / . , , / ,\I II , 'iI I I ,'I I.. . ,

I II I r,I " , A A" ,/ " , I, \ V Y \ n- - -'. . . . . . . ,, ,,,,,

- r ." , \,,,,\"," ' " (," , ,,' , 1 M< :, , ' . ~\ ,,,,\ I"" ,,\ '~!I t t' \ t \

Frequency - Hz

Fig. 31. Radial components at Station 6Wfor two source configurations.

-31-


40/77

10- 4

uIIIIIIN_

UIII"- 1 0 - 5EEICiiic. sc 10- 6. .. _UIIIa.III. . .III~a...

10- 7

- - S D - 2

f1 - - - S D -1~, "'~

, .. , II'" ,I II ', I

/ : ' ' " ~, I, Id, : l~, ~,'\I, V : I, , I ," \ ' \ 1I

~I" '~,

lt 1I fiI,~

I :" j

Frequency - Hz

Fig. 32. Transverse components at Station 6Wfor two source configurations.

a manner similar to that done for CFPVI s,a plot of the distance attenuation param-eter a(f) of the PSD components for eachfrequency is given in Fig. 55 for the Bravo

Event and in Fig. 56 for the 140-ton rowcharge. A comparison of the a(f) of theresultant PSD(f) for the two events is givenin Fig. 57.A plot of a(f)/2 for the resultant PSD

is compared to the CFPV's a(f) in Fig. 58

Figs. 53 and 54 for the same events. In

-32-


41/77

to compare the results of the two spectralanalysis methods. Since there is no re-sultant component of CFPV, an approxi-mation was made by calculating an a (f)from plots of peak velocity for a given fre-quency regardless of which component itoccurred in.

oQ)IIIN_UQ)"- 10-5EE>. ..'t:IIIcQ)"'0c 10-6. .. . . .oQ)c. .III. . .Q)~00-

1 0 - 7

Figure 59 shows the total station energy(L:E) at Station 2BN for the Pre-Gondola IEvents as a function of depth of burst.Figure 60 shows the total energy at Station2BN normalized to depth of burst andplotted vs yield for three yields and sixevents.

SD-2-- - SD-1

A~I f ~ ' "\J;,111II I11 \1 \ III ,j, IIJ' V J N A,I, I\V , .,! .III

Frequency - Hz

Fig. 33. Vertical components at Station 6W for two source configurations.

-33-


42/77

1 0 - 4uQ)onN-Q)"E 1 0 - 5> . . .+-. -neQ)"'00. . .+-u8 - 1 0 - 6n. . .Q)~Q.+-c:0+-::IonQ)c : : :

1 0 - 7

--SO-2/\ - - - SO-l.r-......1 ,,--' ,

I ' ,I ,/ IIIIIdI ,I ,II ,,,,,,\ ' J\ '1 { I,,, I, Y I ~" \ ' I\~ ~, \ l. ,ul,. .I ,,,,~I

Frequency - Hz

Fig. 34. Resultant power spectral densities at Station 6Wfor two source configurations.

In order to give the reader a feeling forthe actual transducer signals, several areshown in Figs. 61 and 62. (The traces

shown have been redrawn for reproductionpurposes from the originals and are notsuitable for analysis.)

-34-


43/77

uQ)II>

cQ)"'00 10-7J:uQ)0-Il>. . .Q)~D..

10-8

A

~ \ -- 50-2\, \ \ --- 50-1,I, ......IIJ ' \ \.. ,\\\,IIII, II\\\' \ 1 .II, , I, \ ~

'"I' I\ , III: I I,\j I III~I~p"I II

~~I ~

Frequency - Hz

Fig. 35. Radial components at Station 2BNfor two source configurations.

-35-


44/77

1 0 - 4 r - - - - - - - - - - - - - - - - - - - - - - - , - ~ - - - - ~ ~ ~ - - ~ ~ - - - - - - ~ - - ~ ~ ~ ,~ ~ - - - - - - ~--SD-2-- -SD-lI

10-5r- -+r+__~~A~A~----~--------------~// , , \ /I \ " n\ I \,/ \

~ E : l O - 6 r - - - - - - - - - - - - - - - - - - ~ - - - - - - ~ _ ; - - - - - - - - - - - - ~ \ - ' j - ' \ ~ \ r _ _ + , ~ \ I r _ _ n ~ ~ I ~ - - - - - - - - - - - - - - - - - - - - - - ~E I ,I:I ~ f , \~I} ~~ i , / V ; I i :.- . : ,\,'" \ "I Ir ,~ ,~~I_ g 10-7 .' : I AU r - - - - - - - - - - - - - - - - - - - - - - - - - + - - - - - - - - - - - - - - - - - - - - - - - - - ~ ~ f t _ - - - - - - - - - - - - - - - - - - - - ~" ,q ~en II '\. . . . ,~ ,~

10-8r-------------------r-------------------+-----------------~

10 - 9 ~ ~ ._ ~10-1 100 101 102

Frequency - Hz

Fig. 36. Transverse components at Station 2BNfor two source configurations.

-36.,.


45/77

1 0 - 5uIIIiii~ UIII

"E 1 0 - 6iiicIII"'00. . .. . . .u8 .iii. . .III~Q..

1 0 - 7

-- 5 D - 2--- 5 D - 1I ", ' ~ I, \ I!1 I \ V .II, 'I, ,I, , I

. . ~,

I ,I ,", j './ \ II I I ,I. . . . . _ \ I I I, I , I " ,, 'i' I 1 '1, , 1 ' \ 1 ' ~I I I:JII III , ' ii ' "\I , "I 'I IIV " "\ "~ I II IIM~~I'I, '

Frequency - Hz

Fig. 37. Vertical components at Station 2BN for two source configurations.

-37-


46/77

10-40C IIII>N. . . . . . . .0C II' "E 10-5

-- 5D-2- -- 5D-1/\I t / ~ \ L I" I II , I, I, / , A I\. . . .,\ I

- : .J i A 'I II II I~', I \: :~ I~I', I \ II : II~ I'IV ,I I\ 1 " ~ I II" I: : '~(III\\ ~ ( ,\

Frequency - Hz

Fig. 38. Resultant power spectral densities at Station 2BNfor two sour-ce configurations.

-38-


47/77

10-1~----------------,------------------r-----------------'

UQ)IIIN. . . . . . .E lu 10-4E ~..._.> -:!:IIIeQ)-00. . .- 1 0 - 5Q)0-III. . .Q)~00...

10-7~----------------~~----------------~~----------------~

5 N10-8~ ~~ ~~~ __~ ~10-1 100 102

Frequency - HzFig. 39. Radial components for Pre-Gondola I Bravo Event at various distances.

-39-


48/77

1 0 - 4 2BNuQ)IIIN-lu 3NE S l._,> -. . . 1 0 - 5IIcQ)"'00. . . .. . .uQ)0-III. . . .Q)~0 1 0 - 6..

10-2r-------------~--~._----------------~------------------,

10-3~--------------4_--~----~----------_+------------------~

4N10-7~--------------~~+_--------------_+~~------------------~5N

10-8L- ~ ~ ~10-1

Frequency - Hz

Fig. 40. Transverse components for Pre-Gondola I Bravo Event at various distances.

-40-


49/77

10-2r-----------------~~--------------._----------~----~

2BN

10-4uC 1 l'"('II-l uE ~- - -> -~ 10-5'"cQ)"D0. .-uC 1 la.'".)~0 10-6.

10-8~ ~ ~ ~10-1

Frequency - Hz

Fig. 41. Vertical components for Pre-Gondola I Bravo Event at various distances.

-41-


50/77

10r----------------.----------------.---------------~

o 10-2)IIIN.--.. 2BNE l uE ~ 3N'-> -.'t:IIIcQ)"'0 10-3o. . . .. . . . .oQ)a.III. . . .Q)~8 .. . . . .co. . . . .:J 10-4IIQ)c o : :

10-6~ ~ ~ ~10-1

Frequency - Hz

Fig. 42. Resultant power spectral densities for Pre-Gondola I Bravo Event atvarious distances.

-42-


51/77

100~----------------~----------------~------------~~

10-7~------------------------~---------------------------~------------------------~10 -1

2BN

UQ)'"N..-..Elu 10 -3 3NE ~- - - -> -:!:'"cQ)'"0C. . .. . . . 10-4Q)a.'". .Q)~0a.

10 -5

Frequency - Hz

Fig. 43. Radial components for Pre-Gondola II 140-ton row charge at variousdistances.

-43-


52/77

100~----------------~----------------~------------~---,

UQ)II)N--.Elu 10-3~. . . _ . .> -. . . .II)cQ)"'0e 3N. . . .. . . . 10-4Q)0-Il). . . .Q)~0Il..

2BN10-2r-----------------;-----~r_--------_+----------------~

10-5~---------4N

10-7~ ~ ~ __~ ~ ~10-1

Freq uenc y - Hz

Fig. 44. Transverse components for Pre-Gondola II 140-ton row charge atvarious distances.

-44-


53/77

100~----------------~--------------~----------------~

2BN

10-2Q)VIN,-...E l oE ~- - - -> -. . . . 10-3IIeQ)'"0C. . .. . . .0Q)0-III. . .Q)~ 10-40..

10-6~ ._ ~ ~10-1

Frequency - Hz

Fig. 45. Vertical components for Pre-Gondola II 140-ton row charge at variousdistances.

-45-


54/77

101.-----------------.-----------------.-----------------~

10-1 2BNoQjenN..-..E l uE :! l. . . . . . .> - 10-2. . . .. ;;s: :Qj'"0E. . . .oQj0..en. . . 10-3~8 .. . . .s ::0. . . .:;enQjex

10-4 4N

5N

10-6L- ~ ~ ~10-1

Frequency - Hz

Fig. 46. Resultant power spectral densities for Pre-Gondola II 140-ton row chargeat various distances.

-46-


55/77

oQ)II>N-lu 10-3E S l

- - -> --I>CQ)-00. . .- 10-4oQ)0-Il>. . .Q)~00..

1 0 - 1 ~ - - - - - - - - - - - - - - - - - - ~ ~ A ~ - - - - - - - - - - - - - - - - r - - - - - - - ~ - - ~ - - - - - '{.\~ . : ' : \~\/ \\ \_ l . . '{\_ . . . . \

2 ,: ' 1 " 1 , , ~ \ \ \ ~10-~--------------~~~'~--~+r"~"------~r-_'------------------~Ii I " \\ I: !' : ' , \ \ \ 1 MI - I \. ,,_i ! " \ \ " . . .\ ! . \ j : " 1 1 l ~'f I ,'\ ~rfl\" 3[: I \. : f . \ I' I )1;\... : Ii A. : I , ! I .j l i i l l

BravoCharlieDeltaAlfa

~ . I I A :1 \ ": i J V 1\ n\ II!' Inr \/ 1

I10-5~----------------_'------------------~+4~--------------~I

10-6~ ~ ~ ~10-1

Frequency - Hz

Fig. 47. Radial components at Station 2BN for four 20-ton events at different depthsof burst,

-47-


56/77

NE.......UE ~. . . . . . . .I>-. . ..;;;cQ)-0o. . .. . .uQ)a.III. . .Q)~o0..

1 0 - 1 ~ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ~ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ~ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -I

/\/ \1 0 - 2 ~ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 4 - - 4 i - - + ' - - - - - - - - - - - - - - - - - - - - - - - - ~ ~ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ~, r - - . !..( i > \ \ \

I :1 0 - 3 ~ - - - - - - - - _ _ ~~lH + ~\~\i~ ~ A _ _ 4- ~/ " f " , \ : ~ 1 , \ \

I' I j \ t : I ' ~ 'I : I. " V ~/ , / / ~ ~ ~ \ { ~ ~ ~ j f \ l l : ~ '. . . .- '. '\ " '~ ;'~ m :II 1 'I.''':1 " " . ' 1 : , \1 0 - 4 ~ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - + ' - - - - - - - - ~ - - - - - - - - - - - - - - - - ~ \ ~ ~ A \ ~ ! ~ 1 J~~HI~~~~::~ ~

/ V ' v J ! I : \ \ \ 1 \ 1 ~ \ i r \I I~. \~ r \ ~V f 1 v ,1 1 I1 IV

1 0 - 5 r - - - - - - - - - ; - ~H~~/~\-I----------------------~'/

)O-6~ ~ ~ ~1 0 - 1

uQ)OIl

BravoChari ieDeltaAlfa

Fig. 48. Transverse components at Station 2BNfor four 20-ton events at differentdepths of burst.

Frequency - Hz

-48-


57/77

10-1~----------------~--------~--------~1----------------~

~II

BravoChari ieDeltaAlfa

1 0 - 2

uQ)II)N,-..E luE ~- - -> -.'t: 1 0 - 3I)c:Q)"'Ce. . .uQ)a.II). . .Q)~0c..

1 0 - 4 v

10-5~ ~ ~ ~10-1

Frequency - Hz

Fig. 49. Vertical components at Station 2BN for four 20-ton events at different depthsof burst.

-49-


58/77

100r------------------r------------------r-r------------.~--~

oQ)IIINE l oS l- - -

,10-5r-----------------~-------------------+------------------~

10-6~ ~~ ~ __~ ~10-1

Frequency - Hz

Fig. 50. Resultant power spectral densities at Station 2BN for four 20-ton events atdifferent depths of bur st.

-50-


59/77

uQ) 10 -2'"NEfuE ~> -. . ..;;eQ)"'C'0~ 10-4. ..uQ)a.'"~Q)~a~0Q)~

10 -61

A II components R esu lta nt (a rithm etic

sum)o Radial (x)o Transverse (y)" Vertical (z)

- Best fit to resultantpeaks

N,.... 10-3~I~. . . .UJW~ 10-4oUJ Al l components Station energy

or LEo Total energy

radial (x) Eo Total energy

transverse (y) E Cb. Total ene rgy

ver tica l (z) E- Best fit to tE

10 100Radius - kmFig. 51. Peak power spectral densitiesplotted against radius for Pre-Gondola I Bravo Event (a comesfrom the equation: PSD (R)= KR-a). p

101r---~~------~----~

1000 10000 100Radius - kmFig. 53. Energy in seismic signal forcomponents and total energy as afunction of radius for Pre-Gondola IBravo Event (a comesfrom the equation:

1 ~E(R) = KR -a).10All components R es ul ta nt (a ri th me ti c

sum)o Rodia l (x)D Transverse (y)" Vertical (z}

- B e st fit to p ea kresultantu~N

Efu 10 - 1~~I> -. . ..;;;c:Q)"'C'0~tQ)a.'"~Q)~0a.~0Q)c..

N,.... 10 -1uQ)' "EUJIN"'Cc0 Sta tio n ene rg y 0J

or ~Eo Tota l energ y

ra dia l (x ) Eo Tota l energ ytransverse (y ) E

t:. T ota l e ne rg yvertica l (z ) E

1 0 - 5 - B es t fit to ~E1 10 100 10000 10 0 1000Radiu s - km Radius - kmFig. 54. Energy in seismic signal forcomponents and total energy as afunction of radius for Pre-Gondola II140-ton row charge

(a comes from the equation:~E(R) = KR -a).

Fig. 52. Peak power spectral densitiesplotted against radius for Pre-Gondola II140-ton row charge(a comes from the equation:PSD (R) = KR =).p

-51-


60/77

5 0--0 z (Verrlccl )x---x x (Radial)0-_0y (Transverse)

2

4Q)c ..o 3Vl

Frequency - HzFig. 55. Slope of power spectral density attenuation as a function of frequency forPre-Gondola I Bravo Event (i , e., slope a comes from the equation:PSD(R, f) = KfR-a(f) .

0----


61/77

eC ll0-oIn

x,,,I ,' "\\\0--0 t PSDattenuation slopex---x CFPV

1.0L---~--~~~~~~----~--~~~~~'_~1--~~~~~~-U0.1 10 100Frequency - Hz

Fig. 58. Comparison of slopes of attenuation of comb filter peak velocities to one-halfof slopes of attenuation of power spectral density resultant for Pre-Gondola II140-ton row charge as a function of frequency.

> -.0-Il>cC ll"'0~.0-UC ll0-Il>~CC ll0- Best fit for Alfa,Chari ie, Delta events. EE. . .ow~

0.5-ton DOB(optimum)

Depth of burst - mFig. 59. Total station energy (~E) atStation 2BN (Pre-Gondola I) show-ing transmitted energy dependenceupon actual depth of burst (Bravodatum is neglected - see text);curve is extrapolated to DOBISfor 40-ton and 0.5-ton events toobtain comparisons for Pre-Gondola II.

Pre-Gondola II(140-ton)

w~ 10-4~~~------~------~ ~

0.1 10 100 1000HE yield - tons

Fig. 60. Total station energy (~E) vsyield after depth of burst effecthas been removed.

-53-


62/77

0.5 mm/sec

:r1.0mm/sec..1... .1.0mm/sec.L

~I sec

Rad ia l Component ., a ll I Iv ln t.PRE-GONDOLA-r-1.0 mm/sec

~ '~'--~~~~~~~~~~~0

. , . . .1.0 mm/sec_ ._-r

1.0mm/sec..1..T1.0 mm/se~...L

Charlie

Bravo

Alfa

Tranlverse Componenll, oillvent.Station 2-NORTHPRE-GONDOLA

Fig. 6Ia. Radial and transverse components at Station 2BNfor Pre-Gondola I (allevents).

r:r1.0mm/sec_L- . - -1.0mm/sec._ .

-r1.0mm/sec....L .

-.-1.0mm/sec...L

Charlie

V er tlc ol C om po ne nt., a ll .ventlStotion 2-N ORTHPRE-GONDOLA I

Fig. 6Ib. Vertical component at Station 2BNfor Pre-Gondola I (all events).

-54-


63/77

Lisee-, -

1.0 mm/seeJ.,_--,'''--'V''I' Charlie

1.0 mm/ sec ---' -"""'1'"" "~JI'WoM""_'__~ =,..,_.___ __-J

I.omm/~~~~~

Bravo

Alfa

DeltaVictor Sum of Compon'nt., all IVlnts

( ".. ultanh)Sla lion 2 -N O RTH

PRE - GO NDO LA I

Fig. 61c. Resultant velocity at Station 2BN for Pre-Gondola I (all events).f---5.0 sec---j STATIO N 2B N RADIALS

0.05I____ -v.r"-"""'~~"'_

1.01Radial 140-ton row charge

0.05I~~ 1.0IE

STATIO N 2 BN TRAN SV ERSES'w'/"M~~'~t-ft\~

5.0 I Transverse Delta0.05I~~1.01

Transverse 140-ton row chargeSTATIO N 2 BN V ERTICALS' Vv I~w' ~wl ' w{ ' l i1yJo ,~~Iw\v .~~~

Vertical S0-1

Vertical 140-ton row chargeFig. 62. Signals recorded at Station 2BN for three different yields (0.5- and 20-tonsingle bursts and 140-ton row charge).

-55-


64/77

Analysis of Results, Discussion, and Interpretation

TIME HISTORYVARIANCES

A comparison of the signals for thePre-Gondola series of events revealed twoprimary facts. The first is that variationsin yield and depth of burst had little effectupon the general appearance of the signalsat any given station except for amplitude.Figures 61 and 62 illustrate this fact verywell. These figures would indicate thatthe source varied only in amplitude andvery little else.The second fact revealed is that in

spite of efforts to instrument sites ofsimilar properties, the signals at thevarious stations bear little resemblance toeach other. For instance, the large earlyvertical arrivals at Station 2BN regularlyconsisted of 7 or 8 cycles of motion ofnearly constant amplitude for all events.At Station 5N for both the Bravo Eventand the 140-ton row charge this portionof the record for the vertical motion con-tained only a single cycle of very highamplitude motion and very little othermotion. Thus when peak velocity isplotted as a function of distance (Figs. 7and 8), Station 5N conai sterrtly appearshigh. When a least squares fit to the datais made, the scatter is large and consid-erable uncertainty exists about projectingthe results to new sites.There is no way to tell whether the

differences in signals between stations isdue to transmission path or local effects,or a combination of these.Because of the unusual motion at

Station 5N during the Bravo Event, whenPNE-11051 was written, Station 5N

figures were neglected and a high slopefor attenuation was given. When the motionat Station 5N from the 140-ton row chargeshowed the same peculiarities, it becameobvious that new slopes should be calculatedfor Bravo. This new value for Q' (Bravo)of 1.54 is cons ider-ahly lower than that re-ported in PNE-ll05 (Q' = 2.45) but showsgood agreement with the Q' for the 140-tonrow charge calculated for the same stations.The attenuation constants of 1.54 and 1.64are lower than expected,6 but Station 5Nstrongly biases these numbers.As was stated, the depth of burst

seemed to affect little except amplitude(Fig. 61). A variation in yield for thesestudies also meant a variation in depth ofburst, thus an even larger difference isnoted in the signals shown in Fig. 62.Part of the difference in generalappearance is due to the high level ofbackground seismic noise in the recordsfrom the SD-1 Event which interfereswith the comparison. Little morethan these qualitative statements can bemade concerning these records but theyare significant statements nevertheless.Although they are not shown, the

records from the SD-2 Event are almostidentical to those of SD-1 except inamplitude. The somewhat surprising re-sult is that amplitudes were conststentlyhigher for SD-2, which was decoupled,than for the coupled shot SD-1. Upon re-flection this is not so surprising. It is awell known fact that an explosively drivenmetal plate will generate a higher pres-sure in a target if it is allowed to bedriven across a gap of inches magnitude

-56-


65/77

as opposed to being initially in contactwith the target.7 The reason is that thegap allows for the potential energy of theexplosive to be changed into the kineticenergy of the metal plate, resulting in ahigher initial pressure in the target.Since the explosive in the SD-2 shot wascontained in a 1/8 -in. thick wall aluminumsphere, 4-3/4 in, smaller in radius thanthe cavity, it is possible that the result-ing higher pressure at the cavity wallcould cause a higher percentage of theexplosive energy to be coupled into theelastic seismic signal.

VELOCITY IN THE FREQUENCYDOMAIN

General BehaviorThe CFPV method does only a partial

job of spectral analysis in that it does afrequency separation of the signal but failsto give any information concerning thetotal energy contained in each frequency.The reason is that a single cycle of motiongives the same value of amplitude as acontinuous sine wave, whereas the twosignals are vastly different in terms ofpower, etc. Because of the greatly varyingtime history for different stations, theCFPV method is unable to clarify to anygreat degree the relationships betweentransmitted signal and received signal.A review of Figs. 9 through 25 illustrates

the important capabilities of the CFPVmethod. A signal usually contains severalsignificant frequencies of motion relatedto the specific modes of wave propagationwith characteristic wave periods. TheCFPV gives an indication of what thesefrequencies are, though only one or two

In general the vertical shows two peaks;one at 2 or 3 Hz due to late arrivals, andthe other at about 8 Hz due to early arrivals.The radial and transverse do not alwaysshow two main frequencies. The higherfrequency is usually smaller in amplitudethan the lower and is sometimes missingentirely. The maximum motion is usuallyfound in either the low-frequency, latearri vals of the radial wave or in the high-frequency. early arrivals of the verticalmotion.The correlation of the CFPV's at the

different stations for the same event isbetter than for the time history signalsthemselves. The general shape of theCFPV's is fairly consistent except atStation 5Nwhere that single cycle pulsegreatly distorts the CFPV of the vertical.This record from Station 5N illustratesone of the main failings of the CFPVmethod.

Range DependenceAs was done for the peak value alone, a

plot of the peak value at each frequency orCFPV vs distance (taken from Figs. 19through 25) may be made and an a be cal-culated from a best fit to the points. Theequation is

CFPV(f) = V (f) = K R-a(f)p f

(These plots are not shown.) When thesea'S are plotted against frequency, thecurve in Fig. 26 is produced. Theory andinvestigations have shown that a shouldbe a function that varies fairly smoothlyand that increases with frequency. 8 Itistherefore surprising to note that the curves

peaks are discernible in most of the figures. in Fig. 26 are far from smooth and, even

-57-


66/77

more surprising, they decrease with anincrease in frequency. This does notappear to be any failing of the CFPV methodsince the same result is obtained by thePSDmethod (see Range Dependence inPower Spectral Density section). Itthere-fore appears to be inherent in the signal.A possibility is that background noise,both seismic and electronic, may be con-tributing factors but no conclusion can bedrawn until further studies are made.

Source ConfigurationGood consistency is shown in the CFPV

method when the same station measuresevents at the same SGZ. This is shown bycomparing the CFPV's of the variouscomponents for Stations 2BNand 6Wonthe SD-1 and SD-2 Events. These areshown in Figs. 11, 12, 13, 16, 17, and 18.There appears to be no significant

change in frequency peaks as the sourcecoupling is changed. But the unusual re-sult of higher amplitudes at nearly allfrequencies for the decoupled event isagain apparent. It is obvious that thiseffect of increase in seismic signalamplitude occurs only for small distancesbetween inner and outer spheres at thesource. Eventually, as this void space isincreased beyond some value, the effectwill be to lower the pressure on the cavitywall thus decreasing the seismic signal.

Yield DependenceOnly two points are available to compareCFPV's at different yields. The point forSD-1 (0.5-ton) is shown in Fig. 9, and thepoint for the 140-ton row charge is shownin Fig. 19. There appears to be a generalshift in the curves to lower frequencies at thehigher yields, though the shift is small,

being only about 30 to 500/0or a factor of280 in yield. Such a small shift may berelated to the actual depth of burst, andone should be careful in trying to extendthis to the completely contained case.The amplitudes are different by approx-imately X100. This is only a crude esti-

mate of the ratio since an accurate one isnot possible due to the shapes of thecurves. A scaling factor of {3= 0.8 isgiven by the equation:

or

A1/AO=(WO/w 1){3100 = (280){3A =kWO.8=kW{3

where W is the yield.

POWERSPECTRAL DENSITYANDSEISMICENERGY

General CharacteristicsThe PSD's for all stations analyzed are

much more consistent than CFPV's. Thisis to be expected since the PSD method isan averaging process, and the problemsexemplified by Station 5Nare overcome.Since it is an averaging technique, the

length of record analyzed becomes im-portant because a single pulse will give asmaller PSD(f) for a longer length ofrecord analyzed. The total energy, how-ever, is independent of the length of record.For this reason the two following stepswere taken to insure consistency. Thetotal length (in seconds) of record wasmade the same when possible so thatvisual comparisons of PSD's would be ob-jectively correct. This length of recordwas 32 sec for all stations and all eventsexcept for Stations 2BNand 6Won the

-58-


67/77

SD-1 and SD-2 Events. The latter recordswere only 24 sec long. Numerical calcu-lations were made, whenever a choice waspossible, on integrated PSD or total energyin a signal or station (E or tE).The radial and transverse componentsconsistently show a pronounced peak at

1 to 3 Hz with smaller peaks at 6 to 8 Hz.The vertical also shows a most pronouncedpeak at 1 to 3 Hz and several strong peaksat 6 to 10 Hz.

Source ConfigurationA comparison of the PSD's for the SD-1

and SD-2 Events, component by componentat Stations 2BN and 6W, are shown inFigs. 31 through 38. These show a re-markable repeatability for the two eventswith only the amplitudes varying. Thisvariance is again an increase of signalstrength for a small decoupling factor.Enough has already been said of thisphenomenon and the results of the PSDmethod are consistent with other analysistechniques.Depth of BurstFigures 47 through 50 show the PSD's

component by component at Station 2BNfor the four 20-ton events. These eventsdiffered only in depth of burst. Againthese figures show little variation exceptin amplitude, a fact not unexpected sincethe actual time histories are so similar.When the PSD is integrated over frequencyto obtain E, and E is summed over com-ponents to give the LE received at the re-cording site, a dependence upon actualdepth of burst becomes apparent. Thisis shown graphically in Fig. 59.The datum point for the Bravo Event

shown in Fig. 59 was low, and reasons

for this were discussed at length in PNE-1105.1 Suffice it to say here that there isgood evidence that Bravo vented early andcoupled less efficiently than it should have.For this reason it has been neglected inthe following analysis.The straight line in Fig. 59 is a leastsquares fit to the three points, Delta,Alfa, and Charlie. It is doubtful that thereal curve is a straight line, but threepoints do not give enough clues to warrantany other choice. As a first approximationthen, the linear fit to a semilog plot ismade. For obvious reasons this curvemust flatten out at very deep depths andthus should not be extended to containmentcases.

It is the contention of the author thatthis dependence of total station seismicenergy (LE) on depth of burst is a resultof the difference in time for the expandingcavity to communicate with the atmosphere.This contact with the atmosphere causesa sudden drop in cavity pressure and, ineffect, an end to its capability to generateseismic signals. Thus the longer a posr-tive pressure (i , e., greater than over-burden) exists in the cavity, the longerwill be the time during which cavity growthand motion can radiate energy into theearth and hence the greater the energycoupled into the seismic signals.By extending the straight line in Fig. 59

one is able to read directly the value of Eat Station 2BN for 20 tons at any depth ofburst within the limits of a few meters tonearly complete containment.

Yield DependenceThe variation of PSD with yield for the

components and resultants at Station 2BNis shown graphically in Figs. 27 through

-59-


68/77

30. These figures illustrate the goodreproducibili ty of PSD for three differentevents.In order to understand the yield depen-

dence of the value of L E for the crateringcase, the depth of burst effect must beremoved. This is done by using the re-sults of the previous section and the curvefrom Fig. 59. The procedure is as follows.Values of L E for all events are normalizedto some standard depth (in this case the0.5-ton optimum depth) by multiplying themeasured L E by the ratio of the 0.5 -tonpoint to the actual depth of burst point.This new value or L E I is the L E whichwould result if the event were detonated atthe shallower depth. The L E I is inde-pendent of depth of burst within the limitsmentioned in the previous section. If themeasurement were made at Station 2BNfor Pre-Gondola events this L E I wouldalso be independent of range. A plot ofsuch events would give the yield-Edependence.The calculated values for the 20-ton

events are:L E A = 16.9X 10-2(7.7/151)= 8.65 X 10-3 (Alfa Event)

LEC = 6.34 X 10-2 (7.7/64= 7.62 X 10-3 (Charlie Event)

LED = 20.3 X 10-2 (7.7/214)= 7.8 X 10-3 (Delta Event)

To calculate for the row charge (three20-ton and two 40-ton charges), eachcharge must be treated separately andmultiplied by its proportion of the total.Thus

-2= 6.64 X 10 (140-ton event)

These numbers are plotted in Fig. 60along with L E for the SD-1 Event, and astraight line fit to the points is made (forSD-1, L E = L E I ) . The slope of thisline is a = 1.11. If all of the variablesexcept yield had truly been normalized out,a slope of 1.0 or less would be expected.Considering the few numbers of pointsused in the depth of burst curve, thisclose agreement is quite good. It isreasonably safe to conclude that the totalenergy scales as the first power of yieldin this range, and that the difference (be-tween 1.0 and 1.11) lies in the inaccuracyof the curve in Fig. 59.

Range DependenceA comparison of the separate com-

ponents at all stations is shown in Figs. 39to 42 for the Bravo Event and in Figs. 43to 46 for the 140-ton row charge. Agree-ment of the curves as a function of rangeis fair for Bravo and better for the 140-ton event. The best consistency is in theresultant shown in Fig. 46.A plot of the peak energy density

(PSD(f) ) for each component and re-maxsultant vs range is shown in Fig. 51 forBravo and in Fig. 52 for the 140-ton rowcharge. The lines are best fits to the peakresultant values. The attenuation slopesof 4.01 and 3.71 are in fair agreementwith each other.An excellent agreement results between

the two events when E and L E are plotted(Figs. 53 and 54). The two attenuationslopes for L E are a = 3.55 and a = 3.53for the two events (see the next section fora discussion of these values).Attenuation of the energy in each fre-

quency band is not made very clear by

-60-


69/77

PSD results. The attenuation slope foreach frequency is shown in Fig. 55 forBravo and in Fig. 56 for the 140-ton eventfor three components. The resultants fromthe two events are compared in Fig. 57.Although there is good agreement be-tween events it is still not clear why a(f)

is not a simple function of frequency andwhy a decreases as frequency increases.8

Comparison of Analysis MethodsIt is clear from the above discussion

that the CFPV method reveals more aboutthe nature of the incoming seismic signalthan a simple peak velocity determination.Likewise the PSD method reveals evenmore information than CFPV, and the re-sults appear to be more consistent withthe mean deviation of various stations beingsmaller for this latter method.An interesting and encouraging result

is obtained by comparing L :E slopes topeak velocity slopes. Let us reason asfollows: (1) the peak velocity, on theaverage (i.e., for many stations and events)will be proportional to the signal attenua-tion, and (2) average energy is proportional

to average velocity squared. It thenfollows that the slope of the peak velocitycurve should be one half the slope of theenergy curve. From the foregoing:

1/2a (total energy, Bravo) = 1.761/2 a (total energy, 140-ton) = 1.76

a (peak velocity, Bravo) = 1.54a (peak velocity, 140-ton) = 1.64a (peak velocity, average) ';;:'1.6

Other investigators have found a'S for peakvelocity ranging from 1.5 to 2.0 with arather generally accepted average ofabout 1.8.1,6,8 Thus the a of 3.55 forenergy attenuation seems to be an excel-lent number.A comparison of 1/2 a(f) from the re-

sultant PSD(f) and a for the CFPV on the140-ton event shows good agreement atnearly all frequencies (Fig. 58). Sincethese methods are different in approach,the results support the validity of eachmethod in that they accurately describethe content of the signal. They do not ex-plain why the high frequencies attenuate(apparently) more slowly than the low fre-quencies.

Concl usionsFrom the foregoing the following con-

elusions may be made:1. Peak velocity, CFPV, and PSD are

all valuable tools provided one under-stands their limitations and applicability.However, for consistency in describingthe transmitted signal in spite of stationdifferences, PSD is the best of the threemethods. In fact, its good consistency incorrelating with yield suggests twopossibilities:

a. This might be a valuablediagnostic tool to determineyields of underground explosions

b. This technique might be appliedto earthquake records and be thebasis for an earthquake magnitudescale.

(In fact, I might even be so bold as to callthis scale the "Power power earthquakemagnitude scale" -albeit, somewhatfacetiously) .

-61-


70/77

2. The total station energy (LE) ofthe transmitted seismic signal over theranges of 0.5 to 140 tons and probablyhigher varies as the first power of theyield.3. Varying the absolute depth of burst

affects the transmitted seismic signal atcrate ring depths to a marked extent. Interms of energy, a factor of 2 for a smallchange in depth of burst is not unusual andmight be important in selecting a depthfor a future cratering event.4. The total station energy (LE) in a

seismic signal over the intermediaterange for frequencies of 0.5 to 20 Hz goesas the -3.55 power of range

J.E = "E R-3.55.:....J L..J 0

5. The effect of geologic differencesin transmission path and local geologiccolumns causes gross differences in thetime dependent appearance of the groundmotion. These geologic differences arenot as effective in altering PSD or thetotal energy arriving at the station whenthe geologic and geophysical properties ofvar-ious stations are similar. A con-clusion as to what would be the result forlarge differences in transmission pathand local geologic properties cannot bemade on the basis of this work, andfurther study is needed.

References1. D. V. Power, Intermediate Range Ground Motions: Project Pre-Gondola I,

Lawrence Radiation Laboratory, Livermore, Rept. PNE-ll05, May 1967.2. A. J. Eardley, Structural Geology of North America, Chapter 5,(Harper and Row,

NewYork, N. Y., 1962) 2nd ed.3. D. L. Watson, Procedure for Recording Ground Motion Velocity from Under-ground Nuclear Explosions, Nevada Operations Office, Las Vegas, Rept.NVO-1l63-93, August 1967.

4. R. F. Beers, Inc., Investigation of Salmon Ground Motions and Effects, NevadaOperations Office, Las Vegas, Rept. NVO-1163-51, April 1965.

5. R. B. Blackman and J. W. Tukey, The Measurement of Power Spectra, (DoverPublications, NewYork, N. Y., 1958).

6. Environmental Research Corp., Preliminary Interoceanic Canal Ground MotionReport, Nevada Operations Office, Las Vegas, Rept. NVO-1163-117, July 1967.

7. R. S. Bradley, High Pressure Physics and Chemistry, Vol. II(Academic Press,NewYork, N. Y., 1963), p. 228.

8. J. E. White, Seismic Waves; Radiation, Transmission, and Attenuation, Chapter 3,(McGraw-Hill Book co., N. Y., 1965).

-62-


71/77

Appendix AA Discussion of the Transmission

of Seismic Signals and their MeasurementThe most universally used method of

determining the ground motion from anunderground explosion is to measure witha transducer a dynamic coordinate of somefinite part of the earth as a function oftime. Ordinarily the coordinate is dis-placement, velocity, or acceleration. Intheory any two may be obtained from athird by integration or differentiation. Inpractice this is difficult because of inherentnoise, instrument drift, instabilities ofnumerical methods, etc. It is also usuallynot possible to measure all three directlydue to the financial expense and instrumentcapabilities. The usual compromise is touse the best instrument suited for themagnitude of motions and to do the bestone can with analysis techniques available.For the intermediate range (a few to afew hundred kilometers), velocity gaugesare the usual choice.The time-dependent motion that occurs

at any given surface location is the resultof a transmitted signal which has had avery complicated, and for the most partunknown, past. This motion is not onesignal but many, each arriving via aseparate path from the source. Thusthere are direct arrivals traveling ex-clusively in the near surface materialwhile other signals are reflected and re-fracted from the deeper layers of theearth's crust. In addition to various pathsthere are several modes of wave motion,such as compressional, shear, Rayleigh,and Love waves each traveling at different

velocities. Finally, for each path, thereare numerous interfaces and impedancemismatches with attendant reflections andalterations in the signal. Certain geologicand geometric structures, particularlynear the surface, may act as wave guidesor resonating structures. Finally, thesignal which the investigator obtains isnot

pne 1115 Pre-Gondola

Documents

Transcript of pne 1115 Pre-Gondola