RAIM 969 STRUCTURAL RESPONSE OF SNIP STIFFENED PANEL AT EVENT 1/1NINON SCALE(U) DEFENCE RESEARCH ESTUBLISNENT SUFFIELDRALSTON (ALIDERTA) J E SLATER ET AL. JAN 66 DRES-495

WL41LIcLASSIRREDF/ 13/ 19 NL

3-15

10 Hl 2-0

li i. IlIII-.tIIII%

* .. ~ M . .-- -- Mr.

National DWense

WencenatinaleUNCLASSIFIEDI 11 liiiLL" 111 ,IISUFFIELD REPORTif ii 11111111 ff1 111 lliliD iiii i °I 485

OlIC f ILE WEY/

STRUCTURAL RESPONSE / '00a) OF SHIP STIFFENED PANEL

oAT EVENT MINOR SCALE (U)

I

0 by

J.E. Slater, R. Houlston I.

and

D.V. Ritzel DTICECTE

~FEB 241988RPCN No. 0113R D

January 1988

DEFENCE RESEARCH ESTABLISHMENT SUFFIELD, RALSTON. ALBERTA

Gmada", .08l 2 2- 1

; * ; :: : : : : : : ..--.!,w .,. , ..... ,...... ..,A....se.w. .... -.a

UNCI.ASSIFIED

DEFENCE RESEARCII ESTABLISHMENT SUFFIELD

RALSTON ALBERTA

SUFFIELD REPORT NO. 485

STRUCTURAL RESPONSE

OF SHIP STIFFENED PANEL

AT EVENT MINOR SCALE (U)

by

J.E. Slater. R. Houlston

and

D.V. Ritzel

PCN No. 0113R

Task No. DMES-26

t%%

' UNCLASSIFI ED

I, I___ __ __ __ _ __ __ __ _ __ __ __ _

IFOLW -VV'~ I%; Kr .rw -6W R .

UNCLASSIFIED

DEFENCE RESEARCH ESTABLISHMENT SUFFIELDRALSTON ALBERTA

SUFFIELD REPORT NO. 485

STRUCTURAL RESPONSE Accession ForNTIrS GRAI

OF SHIP STIFFENED PANEL TIC TRB

AT EVENT MINOR SCALE (U) UnannouncedJustfleation a.I

by _ _ _ ._ ,Distribution/

J.E. Slater, R. Houlston Availability Codesand Avafl and/or

D.V. Ritzel t Specal

ABSTRACT

Modern computational methods and experimental test procedures are beingdeveloped to determine more accurately the air-blast loading and dynamic response ofnaval ship structures subjected to weapon explosions. Previously, a series of air-blasttrials on full-scale stiffened panels typical of those used in warship superstructures hadbeen conducted for conventional scale weapons (under 454 kg HE) to evaluate new designand damage-assessment techniques. The present report describes the experimental resultsand the associated numerical study of the panel response for a longer duration loadinggenerated by the simulation of a 8 kiloton nuclear weapon surface explosion called MINORSCALE. The panel, exposed to a side-on overpressure blast loading of 345 kPa peakand 200 ms positive duration, experienced only slight permanent deformation of about25 mm with no apparent weld or buckling failure. The measured pressure, strain,acceleration and displacement data correlated well with the computed numerical results,thus verifying the computational methods. W.

5, ,(

UNCLASSIFIED

%, . .% ., ,". . .,-,"," .-'-"- -",."+". "-" " ..." . ." .-"-t"," .,.. , ," " ." " " " ",." ""- ''"' & ,"P " +L %- -%

%-a j - -- -' ' , ' , ., , - . " +, " . " .' ' . . - ° % . -% " . % " + " . " . , " + " % - " -" " " " + " " , + + " " " " + . _ , , m

UNCLASSIFIED

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the ADINA analyses carried out by

Mr. C. DesRochers at MARTEC Ltd. Also special thanks to Messrs. G. Rude

of the Shock and Blast Group, and R. Campbell and D. Saint of the ElectronicDesign and Instrumentation Group at DRES for their assistance with theexperiment; and to Mr. C. Workman for processing the data while working forthe Shock and Blast Group as a COOP student under the supervision of

* Dr. R. Houlston.

Jiii

UNCLASSIFIED

-* -

7 T 0 7 Ka: 17 V.-77- 5-7V

UNCLASSIFlIED

TAbLE OF CONTENTS

Page No.

ABSTRACT ................................................................. i

ACKNOWLEDGEMENTS ..................................................... ii

TABLE OF CONTENTS ...................................................... iLIST O F A BLE .. ... .... ... ... .... ... .... ... ... .... ... ... .... ... ..

LIST OF TABLRES ............................................................ vi

1. INTRODUCTION .......................................... 1I

2. OBJECTIVES.............................................. 3

3. EXPERIMENT DESCRIPTION............................... 3

4. COMPUTATIONAL METHODS.............................. 4

4.1 Blast Loading .......................................... 44.2 Structural Response ..................................... 5

5. PRE -TEST DATA ........................ ................. 7

5.1 DRES Panel Data and Analyses........................... 75.2 Preliminary Predictions .................................. 8

6. EXPERIMENTAL RESULTS ................................ 10

6.1 Blast Loading .......................................... 10 L6.2 Structural Response..................................... 11

7. NUMERICAL RESULTS AND CORRELATIONS ............... 13

7.1 Blast Loading .......................................... 137.2 ini e El men ( A D N A) Resuts M A RT C) .... ... .... . .

7.2 Finite Element (ADINA) Results (MRE)...............13

7.4 Structural Response (UBC) .............................. 15

8. CONCLUDING REMARKS.................................. 15

9. REFERENCES............................................. 17

TABLES -

FIGURES

iv

UNCLASSIFIED

* - 'e, ":~ ~ *, - . * -

-~" 4- ILI.~ . 4 .

UNCLASSI FIED)

4W.

LIST OF TABLES

Table I SUMMARY OF DRES I-liIGHT OF BURST (HOB) TESTdoRESULTS FOR STIFI ENI D) PANEL NO. I

.

14 or 4. e.

% % r A A

UNCLASSIFIED '.

I

LIST OF FIGURESP'

Figure 1 Photograph of Stiffened Panel for DRES Tests

Figure 2 Transducer Locations on Stiffened Panel for MINOR SCALE Trial

Figure 3 a) Stiffened Panel Structure Being Installed on Reinforced .tConcrete Foundation for MINOR SCALE Test

b) Panel as Installed Flush with the Ground with DRESBlast-Gauge Stations at Either End of the Panel

Figure 4 Panel Deformation After Final Shot of DRES HOB Tests

Figure 5 Finite Element Model of the Plate/Beam Structure for a Segmentof the Stiffened Panel

Figure 6 ADINA Results of Displacement versus Static Load for Stiffened Panel

Figure 7 Predicted Pressure-Time History Loading for Event MINOR SCALEat Station 19 Near Site of Stiffened Panel Structure

Figure 8 Preliminary Predictions of Displacement-Time History at

a) Center of Panel (Transducer D 12)

b) Midspan of Beam Stiffener (Transducer D 13)

Figure 9 Preliminary Predictions of Acceleration-Time History at

a) Center of Panel (Accelerometer A 2)

b) Midspan of Beam Stiffener (Accelerometer A 3)

Figure 10 Measured Side-On Pressure-Time Histories on Stiffened PanelSurface (Transducers P 1 - P 3) and Blast Gauge Station ."(Transducer P 8)

Figure 11 Averaged Pressure Loading from Data Given in Figure 10 for FiniteElement Analysis I

Figure 12 Measured Displacement-Time Histories at Panel Center (D 12) andBeam Midspan (D 13)

vi

UNCLASSIFIED

" " " . # ." ." ° -'%' # -. °." '" "g 'd"%-'"'° ,€' ''° " - " " " ' - "

".° °''" " '.' °" " # ." °" .° " ,. %' •'. °" ." °'.'" , . . % -'ft

," ',"- , . , " '. ,','_. ,'r ."- - "-,"-;-,",'- ',. :- '-.' . .'.; " ,".',¢" - '- ' - ,' °' .? ;- - . ,'- , ,",'- , ,' , ,'- - '. ", , ".5.

UNCLASSIFIED

LIST OF FIGURES (Cont'd)

Figure 13 a) Post-Shot Photograph of Stiffened Panel Structure After MINORSCALE

b) Final Deformation of Panel Surface

Figure 14 Measured Acceleration-Time Histories on Stiffened Panel

Figure 15 Measured Acceleration-Time Histories on Stiffened Panel

Figure 16 Measured Displacement-Time Histories from Transducers D 12 andD 13, and Double Integration of Accelerometer Signals A 2 - A 5

Figure 17 Measured Strain-Time Histories from Strain Gauges on Top andBottom Surfaces of Panel

Figure 18 Measured Strain-Time Histories from Strain Gauges at Varic,,sLocations on the Panel

Figure 19 Measured Strain-Time Histories from Strain Gauges on Front andBack Edges of the Panel

Figure 20 Refined Finite Element Model of the Plate/Beam Stiffened PanelSegment for Final Dynamic Analysis (MARTEC)

Figure 21 Final Predictions of Displacement-Time History ata) Center of Panel (Transducer D 12)

b) Midspan of Beam Stiffener (Transducer D 13)

Figure 22 Final Predictions of Acceleration-Time History at

a) Center of Panel (Accelerometer A 2)

b) Midspan of Beam Stiffener (Accelerometer A 3)

Figure 23 Comparison of Finite Element Predictions with the Measured andIntegrated Motion-Time History at Center of Panel (TransducersA 2 and D 12)

\.ii

UNCLASSIFIED

10

UNCLASSIFIED

LIST OF FIGURES (Cont'd)

Figure 24 Comparison of Finite Element Predictions with the Measured andIntegrated Motion-Time History at Midspan of Beam Stiffener(Transducer A 3 and D 13)

Figure 25 DRES Finite Element Model using Quadrilateral Element

Figure 26 Averaged Pressure Loading from P 1 and P 3 for DRES FiniteElement Analysis

Figure 27 Comparison of Finite Element ADINA Results with MeasuredDisplacement-Time History at

a) Center of Panel (Transducer D 12)

b) Midspan of Beam Stiffener (Transducer D 13)

Figure 28 Comparison of Finite Element ADINA Results with MeasuredStrain-Time Histories on Panel near Beam Midspan

Figure 29 Comparison of Finite Element ADINA Results with MeasuredStrain-Time Histories on Panel near Clamped Boundary

Figure 30 Comparison of Finite Element ADINA Results with MeasuredStrain-Time Histories on Panel near Clamped Boundary Equivalentto Gauge Locations S 18 and S 26

Figure 31 a) Finite Element Equivalent Beam Model of Stiffened Panel forFENTAB

b) Finite Strip Model of Stiffened Panel

Figure 32 Comparison of UBC Numerical Predictions with ADINA Results andMeasured Displacement-Time History ata) Panel Center (Transducer D 12) Using Finite Strip Method

b) Beam Midspan (Transducer D 13) Using Finite Strip Method

c) Beam Midspan (Transducer D 13) Using Finite Element

Equivalent Beam Method FENTAB

Figure 33 Comparison of the Finite Strip Predictions with ADINA Results andMeasured Displacement-Time History at

a) Panel Center (Transducer D 12)

b) Beam Midspan (Transducer D 13)

viii

UNCLASSIFIED

,]I

I

UNCLASSIFIED

DEFENCE RESEARCH ESTABLISHMENT SUFFIELDRALSTON ALBERTA

SUFFIELD REPORT NO. 485

STRUCTURAL RESPONSE

OF SHIP STIFFENED PANELAT EVENT MINOR SCALE (U)

by

J.E. Slater, R. Houlstonand

D.V. Ritzel

1. INTRODUCTION

a. In the design of naval ships, a capability to assess the dynamic response and damage

of structures subjected to impulsive loads from external or internal air-blasts andunderwater explosions is required not only for the purpose of blast-hardening the design,but also for accurate prediction of the mechanism of failure and the extent of damageonce design thresholds are exceeded. For these types of explosions, the vulnerability andcombat readiness of the complete ship structure must be assessed. The Canadian DNDcurrently has procedures for design of naval ship structures against blast and shock loadingsome of which are based on theoretical and empirical approaches used for decades. TheDefence Research Establishment Suffield (DRES) is developing modern computationalmethods and experimental test procedures for more accurate determination of blast and

shock loading and structural response in order to upgrade and facilitate design analyses.

UNCIASSIFIED

.~ %o

S U ,,a a, -S,.. - .: ? t , *. , , .. -., . .,..,,- -,, -,,,a , aA A -- P

'p,

I

UNCLASSIFIED /2

Metal plates and panels are fundamental elements in many military structures;

particularly ships which are composed primarily of stiffened metal panels that make upthe hull, decks, bulkheads and superstructure. The panels must be designed to withstandair-blasts from either conventional or nuclear weapon explosions. In regard to this

subject, DRES has been tasked by the Director Maritime Engineering and Maintenance(DMEM) and Director Maritime Engineering Support (DMES) to conduct a series of

air-blast trials on full-scale stiffened panels typical of those used in warship superstructuresin order to evaluate new design and damage-assessment techniques being developed as

part of the Shock and Blast research program at DRES (Project Number 011 SA -"Methods for Warship Design to Withstand Blast and Shock"). This research

complements and extends DRES's previous experimental and analytical studies on the

effects of air blast on masts and antennas, which were conducted in the 1960's to mid 1970's.

The panel tests at the DRES Height-of-Burst (HOB) experimental range facilitywere limited to blasts from conventional scale weapons (under 454 kg HE) giving loads

in the impulsive realm with relatively short time duration. The experimental data froma series of tests on square plates, and the correlation with numerical results from the finite

element code ADINA have been presented in References I - 6, and a comprehensivesummary of the experimental results from tests on a stiffened panel is given in Reference 7.

The present report describes a study of the response of stiffened panel structuresto longer duration loading typical of that generated by nuclear weapon explosions. On

27 June 1985, the U.S. Defense Nuclear Agency (DNA) conducted a ground burstsimulation of an 8 kiloton nuclear explosion called MINOR SCALE at their Permanent

High Explosion Test Site (PHETS), White Sands Missile Range (WSMR), NewMexico. For that test, an experiment was fielded to measure the response and damage

of a full-scale stiffened panel subjected to a blast loading of 345 kPa (50 psi) peakoverpressure and 200 ms positive duration. The panel survived the blast loading,experiencing only slight permanent deformation of about 25 mm. The test results andpreliminary numerical correlations were reported at the MINOR SCALE symposium held

in February 1986 [8 1. The present report contains a comprehensive presentation of theexperimental data, a description of the computational methods, and correlation of the ,,

lcI data and numerical results.a'i

U NCI .\SSIFIED

t% % N%

.'

UNCLASSI FIED /3 -'

4,.

2. OBJECTIVES,.-

To fill the technology gap on the response of ship stiffened panel structur':s to longduration air blasts, a panel was fielded at the 8-kt nuclear ground-burst simulation MINORSCALE. The project was undertaken in response to a requirement by DMES under TaskDMES 26. The objectives of the project \were as follows:

I. Conduct measurements of the blast loading and structural responseon a full-scale ship's stiffened panel subjected to a long duration,air-blast overpressure pulse of sufficient strength to producepermanent plastic deformation.

2. Establish the pattern and mechanism of failure and the loadingfunction necessary to cause failure.

3. Correlate the loading and response measurements with design analysiscomputational methods.

Thus, with results and insight gained from the experimental data and application of the -

theoretical and numerical analyses, the overall objective is to develop and validate newdesign and damage-assessment techniques for ship panel structures subjected to longduration air-blast loading typical of nuclear weapon explosions.

3. EXPERIMENT DESCRIPTION

At event MINOR SCALE, 4.8 kilotons of Ammonium Nitrate-Fuel Oil (ANFO)explosive material in a hemispherical fibreglass container was detonated to generate theshock overpressure loading equivalent to an 8 kiloton nuclear explosion. The full-scalestiffened steel panel fielded at MINOR SCALE was identical in structural design to thepanels tested at DRES (see Figure 1 ). The effective size of the Tee-beam stiffened panelwas 2.44 m x 4.57 m with 6.35 mm plating (see Figure 2). The panel was mounted flushwith the ground surface at a radial location 278 m from ground zero to provide a blastloading of 345 kPa peak side-on overpressure and 200 ms positive duration. \Vith203 x 203 x 50 mm sections welded to the 152 x 152 mm box-beam perimeter structure, -5-

the panel was embedded in a reinforced concrete foundation to simulate rigidly clamp:dboundary conditions. Photographs of the foundation And panel site during and after

UNCLASSI FIED

,4.

• €, g' . .. . -. ,- .,. .-. ." . • o % ' . .. . % -. %.l%

, _% % ' ' % "• .. °. . . ... '' " I r .-% .,..

. .' "% " " " _ _ _ _ -%- -, .- = . = 1

444"4 .4 d ,g r u

S.

UNCLASSIFIED /4

the panel installation are shown in Figure 3. Access to the underside of the panel viaa hatch cover and crawl space was provided to allow for attachment of transducer sensorsand hook-up of cables.

As shown in Figure 2, the panel was instrumented at various locations withtransducers for measuring the side-on overpressure, the resulting vertical displacementand acceleration of the panel, and the strain on the panel surfaces. The strain wasmonitored 40 mm from the beam and outer boundary edges. In addition to the panelinstrumentation, two DRES blast gauges [9, 10] were located at either end of the panel( Figure 3(b)) to measure the free- field blast-wave characteristics: static and stagnationoverpressure, shock speed, and density. The transducer analog signals from the panelvere cabled to an instrumentation bunker, and recorded on magnetic tape for laterdigitization and data processing. Following the blast, the panel was examined andsurveyed to determine the shape of the deformed surface and the extent of the damage.

4. COMPUTATIONAL METHODS

An important objective of the project was to correlate the blast loading andstructural response measurements with numerical predictions. This section will describe

the various computational methods that have been developed or are available at DRES.

4.1 Blast Loading

For the blast loading at MINOR SCALE, the location and orientation of the panelflush with the ground reduced the loading to a simple side-on overpressure of 345 kPapeak and 200 ms positive duration. The loading was essentially spatially uniform overthe panel surface due to the long duration of the blast wave except for the initial passageof the shock front across the panel.

There are various methods available for describing the time history of the side-onoverpressure. Baker et al. [11, 12] provide simplified methods for representing thepressure time-history. For the positive phase only, a simple exponential equation of theforim

P(t ) l', '-,"e, (1)

UNCLASSIFIED

id % . .% . , ,% . % ,% • • % ,. . -. , % % .% . % , . ,, " . . . . ,. , , , , % % % .

%

au

UNCLASSI FI ED /5

where P, is the peak overpressure, b. the decay constant, and T, the positive durationis generally adequate. As an improvement, the modified Friedlander equation

P(t) = P,(l , T,,) e'-"',I (2)

has been developed [ 13 ]. This functional form gives a reasonable fit to experimentaldata for the positive phase and the negative phase up to a time t = 1.5 T,.

For structural response analysis using finite element codes, it is advantageous tohave a computerized procedure for automatically generating pressure loads resulting froma specified explosion. MARTEC Ltd. computerized Glasstone's time functionaldescriptions [ 14] of air-blast loading on simple 2-D and 3-D structures, including flat %panels, vertical walls, semicylindrical arches and closed 'box-like' objects. Tabulatedvalues of the dimensionless air-blast parameters (except for the decay constant b) inEquation (2) were obtained from Reference 15. The data for the decay constant wereextracted from Reference 11. The air-blast loading program developed by MARTECis described in Reference 16.

More sophisticated computerized methods that have been developed for blast loadingpredictions include the UVIC Air Blast Compendium [ 171, the Flux Corrected Transport(FCT) Method [18] and the HULL code [191. The Compendium generates timefunctional loading from curve-fits to experimental blast data using the Random ChoiceMethod (RCM) [201. The FCT and HULL codes determine the blast loading by solving Ithe theoretical equations describing the unsteady compressible inviscid fluid flow.

Results from these various methods will be presented in the next section.Comparisons with the experimental results are given in Section 7.

4.2 Structural Response

Numerous analytical and numerical methods have been developed to analyze thestatic and dynamic response of structures subject to various types of loadings. Various

Wefinite difference and finite element algorithms have been coded for predicting the staticand dynamic, elasto-plastic, large deflection behaviour of beams, plates and stiffenedpanels [21, 22]. For linear and nonlinear response analyses, the finite clement codes

UNCI-ASSII lED

.,.

5--. V. * p* ~~ 5~S* * S S S

' ., . , • - -. ,* .• . % - " ". ' . - , . % % % % % - . . - , " - *.- . . - . _ ".% * .

Ip

UNCLASSIFIED /6

VAST [231 and ADINA [241 have been used at DRES. The code VAST, a linear codedeveloped by MARTEC Ltd., has an extensive finite element library, pre- and post-processors for modelling and plotting purposes (25], and a multi-levelsubstructuring/superelement dynamic analysis capability. Special programs have beendeveloped to automatically generate the finite element model of a stiffened panel structure[261 and to describe the time history of the pressure loading [ 16 1. Translators [ 27, 28]have also been written to interface ADINA with the suite of VAST programs for anenhanced analysis capability [29]. Gross deformations involving both material andgeometric nonlinearities can be handled by ADINA.

The basic approach in any nonlinear structural analysis is to initially conduct linearanalyses to establish a suitable finite element model representation of the structure anddetermine the elastic response behaviour. Various finite element models can be generatedautomatically using the VAST pre-processing program. Static analyses can then beperformed to refine the mesh. Natural frequencies and corresponding mode shapes canbe calculated using either direct or subspace iteration. Dynamic elastic response analysesis generally performed using direct integration rather than modal superposition becauseof the impulsive nature of air-blast shock loading. Based on the results of the VASTlinear elastic analysis, structural response can then be analyzed with ADINA to predictnonlinear static yielding and deformation, and the dynamic behaviour leading to the finaldeformation.

Other numerical computational methods have also been developed for responseanalyses of structures subject to blast loading. Olson and Anderson at the University

; of British Columbia have developed a finite element code FENTAB [22] for beammodelling, and applied the finite strip technique [30] to simplify the response analysisof stiffened and unstiffened panel structures [31, 32]. These methods reduce thecomputational requirements substantially; thus making them effective for preliminarydesign analyses and predictions. The finite element code DRESDYN, developed byHansen and Heppler at UTIAS has been tailored for response analysis of shell/platestructures exposed to a nuclear blast and thermal environment [33, 34]. The radiantthermal energy effect is represented in the form of a temperature distribution initiallyapplied to the finite element model. Ihe response of the structure subjected to thepressure-time history is then analyzed.

UNCLASSIFIED

a

0 %% %%

P ." % 0 %W

W. WV~r A7CK - , I! W1w.7- V - 7X L .

,up

UNCLASSIFIED /7J

Structural response predictions using the VAST and ADINA finite element programswill be presented in the next section. Final numerical results from ADINA as well asthe UBC finite element and finite strip codes will be compared with experimental datain Section 7.

5. PRETEST DATA

Prior to the MINOR SCALE event, the response and damage of the stiffened panelwere predicted based on experimental data and analyses from the panel tests conducted

at DRES and finite element numerical calculations. The DRES test data were vital forestimating the transducer response levels, understanding the possible modes andmechanisms of failure, and validating of the finite element ADINA code for structural

analyses of the panel structures subjected to air-blast loading.

5.1 DRES Panel Data and Analyses

A series of air-blast response trials of a stiffened panel (shown in Figure 1 ) hadbeen conducted at the DRES Height-of-Burst facility. The charge weight and

stand-off height were varied, as summarized in Table 1, to investigate both the linear

elastic response and the nonlinear plastic behaviour of the panel. The positive durationof the blast-wave was generally less than one third the panel fundamental period (24 ms)

so that the loading was of an impulsive nature.

The first trial (No. 311) was a "twang" test to check procedures andinstrumentation. The next two trials (No. 312 and 313) provided elastic structuralresponse data. Trial No. 313 was just below yield with maximum dynamic strain levelsapproximately twice the static yield value. The fourth trial (No. 314) took the panel

just above yield with measured strain levels of 4000 p and maximum permanentdeformation of about 25 mm. To achieve an even higher blast loading, two 94 kg chargeswere detonated in the fifth trial. This final shot, which provided a peak reflectedoverpressure of 6375 kPa, produced substantial plastic deformation as shown in

Figure 4. Measured permanent deformation at the midspan of the stiffeners was about

90 mm and the center panel deformed an additional 15 mm. Maximum dynamicdeflections and strains were approximately 125 mm and 6200 p, respectively. The largedeflections resulted in substantial membrane stresses being developed in theplating. However, no visible cracking or weld failure was observed.

UNCI.ASSII:LD

IQ '..."','" .,' '.- :" ".". -. .:. '- -"-':'.-.''.- .". "&"."-. " " .' -. -" 7 """¢ & . -'¢-Z .T.%;K -',.,, ,€ ,','7. ,. ,,'%- -% '

UNCLASSIFIED /8

The fundamental natural frequency was observed to be about 41 Hz (period 24 ms)corresponding to the natural resonance of a single clamped panel segment. At eventMINOR SCALE the duration of the shock wave was expected to be about 200 ms, sothat the response of the panel would be in the dynamic realm. Based on the DRES panel

results, particularly trial number 314, it was anticipated that the MINOR SCALE panelwith a peak overpressure of 345 kPa would experience initial plastic deformation of about

50 mm followed by elastic resonance vibrations and a final permanent deflection of the

order of 25 mm. The maximum strains and accelerations were predicted to be about4000 Mi and 2000 G, respectively.

Following the panel trials at DRES, various linear and nonlinear analyses were

conducted to validate the finite element codes VAST and ADINA for structural responseto the blast wave loading. These analyses required tensile tests at quasi-static and lowstrain rates to obtain the mechanical properties of the panel material. The followingvalues of static material properties were measured: Young's Modulus = 207 GPa

(30 x 106 psi); Poisson's Ratio = 0.3; Yield Stress = 310 MPa (45 ksi); TangentModulus = 1230 MPa ( 178 x l01 psi); Ultimate Tensile Stress (at F = 0.2) = 565 MPa(82 ksi). The mass density was taken as 7770 kg/iM3 . For an anticipated strain rate

of 1.2 s-', the dynamic Yield Stress and Ultimate Tensile Stress values were determinedto be 375 MPa (54.4 ksi) and 620 MPa (90 ksi), respectively. The VAST and ADINAstructural results have been correlated with the test data and reported in References I - 5.For the short duration type of blast loading at DRES and for moderate deformation ofthe panel, good agreement was obtained between the predicted and observed panel

motion.

5.2 Preliminary Predictions

Based on the above calculations, linear and nonlinear analyses were conducted to

obtain more accurate predictions of the dynamic response expected at MINOR SCALE.The results of these calculations were used for finalizing the design of the experiment and

for setting the transducer signal conditioning and amplifier gain to give optimum rangefor the transducer sensitivity and recording.

Using the finite clement model sho\\ n iii Figure 5 for a portion of the stiffened

panel, nonlinear static ADINA analyses [351 were performed (see Figure 6). Compared

UNCI.ASSIFIED ',

SR

J. ..C -. , -

- - - -- N Vk-J %X -. b

U NCLASSI: ED /9

to the linear results, the nonlinear geomc ri and material effects (i.e., large dkplaceient "

and plasticity) substantially alter the predicted displacement. The nonlinear stiffening

effect due to membrane stresses is apparent ccn at low levels of loading; and at 210 kPa,yielding of the material becomes significant. Based on the nonlinear material ADINA

static analysis, the final deflection of the pancl for an overpressure of 345 kPa was then

estimated to be about 50 mm with maximunm displacements during the first few cyclesof motion being as large as 100 rm.

p

Calculations were also undertaken to predict the time history of the dynamic

response behaviour of the panel. The expected pressure-time history loading on the panelpredicted by the simplified method (Equation 2) of Baker et al. [ 11, 12 ] and the MARTEC

computerized method [ 161 were used for selecting the site of the panel. More detailedprediction of the expected pressure loading at MINOR SCALE was provided by S -Cubed

Inc. using the HULL code [ 19]. Figure 7 shows the predicted pressure-time history at

a location near the panel site. The amplitude was scaled to obtain a pressure-time curvewith a peak overpressure of 345 kPa. The positive impulse was predicted to be

15.9 MPa-ms.

Using the loading data provided by S-cubed as input to ADINA, the dynamicresponse of the panel was computed for the first 40 ms. The displacement-time histories

predicted at the center of the panel (node 289, transducer D 12) and at the midspan of

the beam (node 303, transducer D 13) are shown in Figure 8. The initial elasto-plastic

deformation due to arrival of the shock front was predicted to be - 75 mm for the panel

and - 50 mm for the beam. This was followed by elastic vibration at 110 Hz (9 imsperiod) which tended to approach final deflection values of about -45 mm for the panel

and -35 mm for the beam.

The acceleration-time histories predicted for accelerometers A 2 and A 3 located

at positions D 12 and D 13, respectively, arc shown in Figure 9. The initial negatie-

acceleration was expected to be about 900 G, with a subsequent positive acceleration about

1600 G for the panel and 1200 G for the beam. The prediction for acceleration A 7 at

node 145 was similar to A 2.

At the time of the calculations, the time history for the components of stress rather

than strain was available from the ADINA output. Because plastic deformation of the

UNC LASSIFIED

% % % %R21 -- , 0 'e%: W X0'J. .P

%C

S s *. . *.*~ ,.-~. ~ .. ,,* C

UNCI ASSIFIED /10

panel \as expected, the stress could not be directly converted to strain. For the purposeof setting-up the strain amplifiers, the ADINA stress predictions and the strainmeasurements from the previous DRES trials were considered quite adequate. Thedynamic stress predictions were converted to strain using Young's Modulus for the elasticportion and Tangent Modulus to account for plastic yielding. Strain levels of the orderof 4000 t were predicted on the upper surface due to combined bending and membraneeffects.

6. EXPERIMENTAL RESULTS

Immediately after the MINOR SCALE Event, preliminary plots of all transducersignals were generated to determine what test data were successful recorded and to provide

dat a for the 'quick -look' report. The panel was examined to assess the extent of damage.The final deformed shape was measured using a precision survey level. Subsequently,the recorded time history data for pressure, displacement, acceleration and strain weredigitized. The computer program SPADE [36] was used to correct the signal baseline,apply the calibration factors, and remove any extraneous noise signals. This programwas also used to integrate the acceleration to obtain velocity and displacement, averagethe pressure-time history traces, and to produce time history plots.

6.1 Blast Loading

Figure 10 shows the pressure-time histories recorded by transducers P 1, P 2 andP 3 located at the panel surface and P 8 at a blast gauge station [9] near the panel. Theinserts show the initial 50 ms of the pressure pulses. These pulses have some unexpected

perturbations which could be attributed to ground surface roughness and boundary layereffects, local anomalies in the shock front, ground/panel surface discontinuity or straingauge cable obstruction near the pressure transducers. The peak overpressure wasapproximately 345 kPa ( 50 psi), with a positive impulse of 15.7 MPa-ms and a positiveduration of 200 ms. At 750 ms, a secondary shock characteristic of blast waves for highexplosives is evident.

As evident from the inserts in Figure 10, there are differences in arrival time ofthe shock front at the various transducer locations. In this figure, zero time is chosenas the average of the arrival times at the four transducers. The differences in arrival times

UNCI ASSIFIED

%. -. * .- . '. - %* - .. . '. '

* % %',p -- P. ' 'p f .. . ... .. . ...* ~% ~ % ~ ~

UNCLASSI [1) /I I

is partially due to the velocity of the shock wave across the panel. The shock wave transit

time from the front to back edge of the panel is about 3 ms. However, correlation of'the shock arrival times from all transducers (reported in Reference 9) shows that there

is evidence that the incident shock front \,as distorted in shape at this site location. Over 1-

a span of the panel site, the shock front had an inflected curvature and was obliquely %incident by as much as 28 degrees from the expected radial propagation. Since there

is no above-ground structures within 100 ni of this site, a possible explanation for thisdisturbance is a reflection from a road and ditch alongside the panel site. The road had

a hard-packed gravel surface causing slightly higher shock velocity and the ditch was about

0.7 m deep by 2 m wide. Both ran parallel to the blast radial line and came within about.,

3 m from the panel site.

Although some aspects of the blast flo\, arrival times appear irregular, these

irregularities should not strongly affect the panel response. Thus, for structural analyses

purposes, the arrival times of the four signals were synchronized and the amplitudes

averaged to generate the representative pressure-time history shown in Figure !1. This

averaged pressure curve was used as the applied load in the final numerical analyses

performed by MARTEC (Section 7.2).

6.2 Structural Response

The displacement-time histories measured by transducers D 12 and D 13 (seeFigure 2 ) for the panel center and beam midspan, respectively, are shown in Figure 12.

The initial deflection is approximately - 53 mm for the panel and - 30 mm for the beam,followed by elastic oscillations about a mean that shows some recovery until the arrival %"

of the second shock. It is important to note that the recovery of the mean deflection .9-

follows a shape similar to the exponential like decay of the blast load amplitude (Figure 11 ).

This in inherently related to the long duration of the MINOR SCALE blast wave. The

final (permanent) deflection is approximately -25 mm for the panel and - 16 mm for

the beam; corresponding closely with the post-shot survey measurements of -25 mm and

- 15 mm for the panel and beam, respectively. A photograph of the panel structure and

a diagram of the final shape of the panel surface (magnified by a factor of 10) are presented

in Figure 13. Values for the permanent deflections at the panel centers and beammidpoints are included in this figure.

UNCLASSIFIEDa'.

-S"f .,i - .p . ." i - P S A

UNCLASSIFIED /12

The accelerations measured by the various accelerometers are presented in Figures 14and 15. The accelerations A I and A 3 at the midspan of the beams are very similar.When the traces of A 2 and A 3 are plotted together the increased motion of the panelcenter compared to the beam midspan can be clearly seen. Accelerations recorded atthe other points on the panel are similar to that at A 2. The combined plots in Figure15 indicate time delays in the various traces due to the transit time of the shock frontacross the panel (about 3 ms) as well as the additional time difference caused by the localoblique flow anomaly in the shock front previously discussed.

By integrating the acceleration twice, the displacement motion is obtained.Figure 16 shows the resulting displacement-time histories at various points on the panel.

In the upper two plots for the panel center and beam midspan, respectively, the integrateddisplacements are compared with the traces measured directly with the displacementtransducers. Excellent agreement is obtained in both cases, confirming the accuracy of

the displacement obtained from double integration of the acceleration. Displacements ,,at the center of the adjacent panel segments given by A 4 and A 5 are similar in levelto A 2. but A 5 exhibits slightly more elastic vibration, being near the panel edge.

Plots of various measured strain-time histories are presented in Figures 17 - 19.No data were available from S 15, S 17, S 21 and S 25 because of gauge failure. Thestrain measurements on the upper and lower surfaces shown in Figure 17 illustrate thetensile and compressive bending strains, and the presence of an in-plane membrane straincorresponding to the difference between the two signals. The strain signals presentedin Figures 18 and 19 show the consistency obtained across the panel, and the relative levels

of the strain normal and tangential to the panel edge about 40 mm from the boundary. Inthe bottom diagram of Figure 19, the initial signals from strain gauges S 18 and S 20 arriveat slightly different times (about 3 ms) due to the transit time of the shock front acrosst he panel. Peak strain levels of the order of 3500 p z were measured by the gauges onthe upper surface due to combined bendin,- and membrane effects. Along the actuallines of attachment of the plating to the beam stiffeners and outer boundary edges, evenhigher levels of strain are anticipated due to pla,ric hinge effects. The initial strain ratecalculated from the plots ranges from 0.76 to 1.9 s- which spans the anticipated valueof 1.2 s' used to determine the material pioperties in the pretest finite element analysis

(,cc Section 5.2).

%,

UNCI \SSIFIlEDI

UNCLASSIFIED / 13

7. NUMERICAL RESULTS AND CORRELATIONS

7.1 Blast Loading

Since the dynamic response of the panel structure depends on the loading function,

the measured pressure-time history was employed for the final numerical calculations.

Although the finite element code can incorporate the transit time of the shock front overthe panel and any other spatial variations, the analysis was simplified by assuming that

the pressure load was applied instantaneously and that it was spatially uniform over thepanel surface. The pressure-time history (shown in Figure I I) was generated by

synchronizing and averaging the amplitudes of the actual measured pressure traces. The

averaged loading data agree well with the HULL code predictions ( Figure 7 ) except forthe perturbations in the initial waveform between 2 to 14 ms. As will be evident by the

following results, the perturbations in the initial waveform do affect the dynamic response

and vibrational characteristics of the panel structure.

7.2 Finite Element (ADINA) Results (MARTEC)

Using the ADINA code and a refined finite element model (Figure 20),MARTEC [ 371 re-calculated the displacement and acceleration at the center of the panel

(node 457, transducers D 12 and A 2) and at the midspan of the beam (node 472,transducers D 13 and A 3) (see Figures 21 and 22, respectively). As shown in Figure 21,

the initial elasto-plastic deformations are now determined to be - 48 mm for the paneland -28 mm for the beam; about 40% smaller than the preliminary predictions (see

Figure 8) based on the HULL code predicted loading. The reduction in maximum

dynamic displacements is a result of the reduced effective loading due to the perturbations

in the actual shock wave between 2 - 14 ms. By projecting the curves to later times,

it appears that the permanent deflections approach about - 25 mm for the panel and

- 15 mm for the beam, corresponding closely with the measured deformations of the panel

-25 mm) and beam ( -16 mm) shown in Figures 12 and 13. The acceleration-time

history for accelerometer A 2 (Figure 22) has a range (-900 G to 1600 G) similar to

the preliminary predictions (see Figure 9), but the acceleration at A 3 on the beam has

a range ( -500 G to 800 G) which is somewhat less than that originally predicted.

Further comparison of the ADINA numerical results with the experimental motlion

UNCLASSIFIED

e-*"_,

",....,-.., -,".-" . ".".-"...':':::; -.':':_';." :'.; :..",:'4-.,Z _,:,:";",: . -.;-.'..% .,..,-% : ¢-.'-. ."-'.. '-., .'- .,.5,.'.7,-"',.;v ,.".'.2 '.'

UNCI .,ASSI I:I ED /14 we'm'p

data shows good agreement (Figures 23 and 24). The measured and predicted acceleration

and displacement at the panel center and beam midspan have similar levels andwa~eform. The initial plastic deformations occur within the first 6 ms of the applied

shock wave loading and peak deflections occur within one period (24 ms). The

experimental data also exhibit some higher frequency vibrational characteristics due to

harmonic vibrations and/or transverse wave vibrations in the plating. Some of theseibrations could also be extraneous noise generated by transducer ringing, and/or vibration.

7.3 Finite Element (ADINA) Results (DRES)

Correlation of the numerical predictions with the experimental data depends of

the accuracy of the finite element model and the loading function. The size of the time

step for numerical integration and the interpolation procedures and convergence criteriaarc also important.

Using the latest version of ADINA (1984), further analyses were conducted byrefining the model, extending the numerical calculations to 1000 ms and producing strain-

time history results for direct correlation with the experimentally measured data. Thenev, finite element model (Figure 25) was developed using the 4-node quadrilateral

plate 'shell element. An extended pressure curve obtained by averaging transducers P 1

and P 3 was used as the pressure loading ( Figure 26).

The excellent correlation of the displacement results for the panel center well beyond

the scond shock is shown in Figure 27. The prediction of too large a final beam deflection

is due to o,,er stiffness of the finite element model for the beam after plastic deformationand the lateral membrane restraint for the remainder of the panel. Modelling one quarter

of the complete panel and refining the model with more elements through the web height

and across the flange would reduce the stiffness and improve the displacement predictions Iduring the recovery phase of the panel dcflection.

The numerical results for strain-time history are determined at the gauss integration

points of the nearest finite elements and then interpolated to strain gauge locations.

Correlation of the strain-time historic,, for a number of the gauges are shown in

I i eurcs 28 30. Numerical results fov S 15 and S 17 are included in Figure 28 een

IOt In11o cperimcntal data were available because of gauge failure at these locations.

UNCI.\SSIFIED

,A

t _ '--, - -

"S

UNCLASSII-lED 115

The difference in start time between the predicted and measured strain-time historic, inFigures 29 and 30, is simply due to the swccp tine of the shock front o\et the panel rclati'Wto the synchronized and averaged time of the loading function used for the analyses.

7.4 Structural Response (UBC)

Numerical results for the panel motions \\ere also obtained using newly developedfinite element and finite strip codes. The finite element equivalent beam and finite stripmodels of the stiffened panel are shown in Figure 31. These calculations, were performed

using the pressure-time loading shown in Figure 26.

The numerical displacement results obtained using these models for the initial 50 insare illustrated in Figure 32 together with the experimental data and the DRES ADINA

calculations. The finite strip calculations were then carried out to 440 ms (see Figure 33 ).The panel center is predicted to exhibit greater elastic vibrations and then a bucklinginstability at 280 ms, referred to as an 'oil-canning' phenomena, after which the platingvibrates in a bowed-up shape relative to the deformed beams. If the buckling instabilityis ignored (since it did not occur), the finite strip results are in good agreement with the

measured data. The effects of damping on the elastic vibrations and the occurrence ofthe 'oil-canning' phenomena warrant further investigation. There are obvious advantages

to using the simplified methods in that computational times are between 10 to 20 minutescompared to 30 hours for the ADINA calculations, so that the predictions are obtainedat substantially reduced cost.

8. CONCLUDING REMARKS'1'

Event MINOR SCALE provided a unique opportunity to measure the responseof the stiffened panel structure for a long duration pressure loading typical of a nuclearweapon explosion. The 345 kPa peak overpressure level with about 200 ins positiveduration exposed the panel to a loading sufficient to cause permanent deformation.

The permanent deformations of 25 mm for the panel and 15 mm for the beamwere less than the pretest predictions. Acceleration levels between 1000 - 2000 G wereabout wlbat were expected. Strain levels up to 3500 W were measured on the upper surface

due to combined tensile bending and membrane effects. Higher strains are likely (butwere not measured) closer to the boundary as a result of the plastic hinging effects.

U NCLASSI FIFD

.

.1 VP% ** %***%%

n7N -4 .% * % -.

UNCLASSIFIED /16 ,

The discrepancy between the experimental data and the pretest predictions wasdirectly related to the loading function used in the analyses. By using the actual measuredpressure loading the dynamic response analysis showed excellent agreement between thenumerica) results and the measured acceleration and displacement data. This illustratesthat in order to examine the details of structural response, it is necessary to use the actualpressure loading. Nevertheless, for the purpose of preliminary design analysis andplanning an experiment, the HULL code prediction was found to be quite adequate ingiving the general features of the pressure loading. ,

The success of the panel experiment at event MINOR SCALE substantiated thetrials procedures, data recording and processing techniques that were used. Thecorrelation of results verified the suitability of the finite element code ADINA for detailedanalysis of structural response, whereas the simplified computational methods based onthe finite element equivalent beam and finite strip techniques have been shown to be suitablefor preliminary design analysis.

The experiment fulfilled the objectives of the task by subjecting a full-scale ship'stiftened panel to a long duration air-blast loading sufficient to cause permanent plastic

deformation, and by establishing the resulting pattern and mechanism of failure. Thegood correlation of the experimental data and numerical results validated the computationalmethods for long duration type of loading function and for structures sustaining limitedamount of permanent deformation. The methods can now be used with confidence in,hip design for carrying-out structural response analyses and damage assessment studies.

Ho~cvcr, for structures experiencing considerably higher plastic deformation anddamage, further tcts and analvse, are rccommended for complete validation of thecorn put ational methods. %

sq'

5.

*5%\SSIF°E

% % % % %

VI.

U NC I. AS'SII I l)'D

REFERENCES

I i.E. Slater and R. fouk ton . -)'~nanilc ,knal\ sI, of licaiw uid Plate,,Subjected to Air-Blast Loadimn&'. Plre. ented at the 3rd Inter. .Modal%

Analysis Conf., Orlando, January 19S5.

2.R. HIuSton and J.E. Slater. "Stilwctur al Response of Panel" Suh iceto Shock Loading"'. Presentcd at the 55th Shock atid VibrationSymposium, Dayton, October 19S-4. P~ublished in the 551h Shock andVibration Bulletin, Part 2, 149 103. June 1985.

3. R. Houlston, I.E. Slater, N. Pegg and C. DesRochers. "On Analysis of'4

Structural Response of Ship Paniels Subjected to Air BlastLoading". Presented at the 5th ADINA Conf., M.I.T., June 1985.Published in the Computer and Structures, Vol. 21, No. 1/2, pp. 273 -- 289,1985.

4. R. Houlston and i.E. Slater. "A Summary of Experimental Re:sults onSquare Plates and Stiffened Panels Subjected to Air Blast Loading".Presented at the 57th Shock and Vibration Symposium, New Orleans,Louisiana, USA, October 1986. Published in the 57th Shock and VibrationBulletin, Part 2, 55 - 67, January 1987.

5.R. HUSton and C.G. DesRochcrs . "Nonlinear Structural Response of'Ship Panels Subjected to Air Blast Loading". Presented at 6th ADINAConference, MIIT, June 1987. Published in Computers~ and Structures,Vol. 26, No. 1/2. pp. 1 - 15, 1987.

6. R. Houlston, 1). Ritzel, I.E. Slater, G. Rude and R.T. Schrnitke."- Blast Experiments on Square Plates". Suffield \Ien. orandurn

No. 1143. lanuLarN 1986. ( U.

7. R. Houlston, D. Ritzel, I.E. Slater and (I. Rude. "Air-Blast Experimentson DRES Stiffened Panels I " Suffield Memiorandumn No. No. 1214,in review. (U )

L N(I[ .\S-)[ 1 11:1

%I

UNCIASSIIFIED /18

REFERENCES (Cont'd)

8. J.E. Slater. "Ship Stiffened Panel Structure". DNA MINOR SCALESymposium, February 1986.

9. D.V. Ritzel. "Measurements from DRES Blast Gauge Stations: West

Radial". DNA MINOR SCALE Symposium, February 1986.

10. D.V. Ritzel. "The DRES Blast-Gauge Station". Presented at the NinthInternational Symposium on the Military Applications of Blast Simulation

(MABS 9), England, September 1985.

11. W.E. Baker, J.J. Kulesz, P.S. Westine, P.A. Cox and J.S. Wilbeck.

"A Manual for the Prediction of Blast and Fragment Loading onStructures". U.S. Dept. of Energy Report, DOE/TIC-11268, November

1980.

12. W.E. Baker, P.A. Cox, P.S. Westine and J.J. Kulesz. Explosion Hazardsand Evaluation. Fundamental Studies in Engineering 5, Elsevier Scientific

Publishing Company, 1983.

13. W.E. Baker. Explosions in Air. University of Texas Press and Southwest

Research Institute, 1973.

14. S. Glasstone and P.S. Dolan. The Effects of Nuclear Weapons. Third

Edition, The United States Dept. of Defence and the U.S. Dept. of Energy,

1977.

15. Engineering Design Handbook - Explosions in Air, Part One.

Headquarters, U.S. Army Material Command, AMC Pamphlet,

AMCP-706-181, 15 July 1974.

16. C.G. DesRochcrs and M.E. Norw ood. "A Computerized Procedure for

l)efining Air Blast Loading on Structural Elements: User's Manual".MARTFIC Ltd., Halifax, N.5.. l)RLS Contract Report No. 28/ 86, March ',

1986.

UNCI ASSII ILD

...................... ,-

UNCL-ASSII IEl) "19

REFLERENCES (Cont'd)

17. J .M. DewkeN and D.J . Mc \l1iini "Compendiumi of Blast Wa~c 0Properties". F'inal Report, Li\ e1~t\ of Victoria, B.C'.,DR ESConitract

Report No. 17 /87. February 198".

18S. T. Otvos. "General Tw- l~rnlerhonal Cornpressiblc 1:1lo\ Calculations

Using SPLIT2D) U~ser'N, Guid". Version 1 .3, ('0ombuLStion Dynami~c"

Ltd., November 1986.

19. M.A. Fry. "The HULL Hydrod\11ami1CS Computer Code". AFWkL-TR-

76-1 83, Air Force WVeapons 1ahmuiaory, September 1976.

20. S.C.M. Lau and JJ.l. Giottlieb. "N Urerical Reconstruction of Part of an s

Actual Blast -Wave Hlow Field to Ae\rkc with Available F\perimecltal Data".

UTIAS Technical Note No. 251, LUniversitv of Toronto Institute for

Aerospace Studies, Downsvie\w, Ontario, August 1984.

21. T.R. Stagliano and L.J. Mente. "Large Deflection, Elastic-Plastic

Dynamic Structural Response of Beams and Stiffened or Unstiffened

Panels - A Comparison of Finite Element, Finite Difference and Model 0Solutions". Nonlinear Finite Element Analysis and Adina, Proc. ADWA

Conference, 1979.

22. B.R. Folz. "Documentation for ie Program FENTAB, Finite Element

Nonlinear Transient Analysis of Beanis., Version 1 .0,. D~epartmnent of Civil

Engineering, University of British CoIlumbia, Vancouver, B.C., May 1986.

23. "Vibration and Strength Analy~s, Program ( V'AST ) U ser's Manual".

Version No. 04, MARTEC Ltd., Halifax, N.S.. January 1986.

24. "ADINA -A Finite Element Program for Automatic Dyvnamic Incremental.

Nonlinear Analysis". Report AL 84-I1, AI)IN.A Engineering Inc., Mass.

USA, December 1984.

ULNCI ASSIlIL -

I-

UNCLASSIFIED /20

REFERENCES (Cont'd)

25. "Vibration and Strength Analysis Graphics Program (VAST-G): User's

Manual". MARTEC Ltd., Halifax, N.S., 1986.

26. C.G. Des Rochers, D.F. Fanning and M.E. Norwood. "Computer Codes

for Structural Response to Air Blast: User's Manual". MARTEC Ltd.,

Halifax, N.S., DRES Contract Report No. 35/86, March 1986.

27. C.G. DesRochers and M.E. Norwood. "VAST/ADINA Interface

Program (ADIDAT): User's Manual". MARTEC Ltd., Halifax, N.S.,

July 1983.

28. C.G. DesRochers and M.E. Norwood. "VAST/ADINA Graphics

Interface Program (ADIPOS): Final Report". MARTEC Ltd., Halifax,N.S., October 1983.

29. C.G. Des Rochers and M.E. Norwood. "Analysis of Ship Stiffened PanelStructures Subjected to Air Blast Loading: Final Report". MARTEC Ltd.,

Halifax, N.S., DRES Contract Report No. 27/86, March 1986.

30. Y.K. Cheung. Finite Strip Method in Structural Analysis. Pergamon

Press, 1976.

31. R.M. Khalil, M.D. Olson and D.L. Anderson. "Large Deflection,

Elastic-Plastic Dynamic Response of Air-Blast Loaded Plate Structures bythe Finite Strip Method". Structural Research Series Rept. No. 33, Dept.of Civil Engineering, University of British Columbia, Vancouver, B.C.,

January 1987.

32. M.D. Olson and D.L. Andcrsoii. "Modelling of Structural Response to

Air Blast Phase II". Final Report for Contract No. OISG.97702-R-4-8568, )epartment of Civil Eneinecring, University of British Columbia,

\"aticou cr, 3.(., lainuarv 1987.

%L! N(I.ASI FIIF ) I

*-... -. ~

-t P. T F.- . ..I .F....--'--

.-" - -

UNCLASSIFIED /21

REFERENCES (Cont'd)

33. J.S. Hansen and G.R. Heppler. D')cfence Research Establishment Suffield

Dynamic Code (DRESDYN) Users' Manual". UTIAS, Toronto, DRESContract Report No. 30/ 85, September 1984.

34. J.S. Hansen and G.R. Heppler. "Development of a Finite ElementCapability for the Analysis of Structural Systems Subjected to a Nuclear

Blast Environment - Phase 11 Beams and Composite Material Shells".Final Report Submitted to Defence Research Establishment Suffield under

Contract No. OISG.97702-R-4-8710, UTIAS, Toronto, November 1986.

35. C.G. DesRochers and M.E. Norwood. "Finite Element Analysis of

Stiffened Panels Subjected to Air Blast Loading". MARTEC Ltd., Halifax,N.S., DRES Contract Report No. 55/85, April 1985.

36. R. Houlston and J.E. Slater. "SPADE - An Interactive Computer

Program for Signal Processing And Data Enhancement of Air-Blast

Structural Response Measurements (U)". Suffield MemorandumNo. 1123, March 1986. UNCLASSIFIED.

37. C.G. DesRochers and M.E. Norwood. "Finite Element Analysis ofStiffened Panel Structure Subjected to Three Air Blast Loadings".

MARTEC Ltd., Halifax, N.S., DRES Contract Report No. 26/86, March1986.

UNCIASSITI-I)

, '

A.'b'~£",,4£",2""',. d',2r j'' 2 .,2eJ2 .? .? 2 2',£ .2'''''''._.2z .2'''. ''¢'.'._',£'_'_''e '..

UNCLASSIFIED

Table I .

SUMMARY OF DRES HEIGHT OF BURST (HOB)

TEST RESULTS FOR STIFFENED PANEL NO. 1

CHARGE STAND- PEAK CE -TIL MASS OFF PRESSURE DRTO IMUS:RATION* STRAIN REMARKS t

NUMBER (kg) (in) (kPa) (ins (kWa-ms) (G) (E

311 3.6 21.3 17 9.3 55 + + ELASTIC

312 29 15.2 83 13.3 300 370 1300 ELASTIC

313 94 12.8 240 10.0 710 880 300 YIT ELD"

314 94 10.1 450 8.2 980 1650 4000 JUS ABOVFYIELD

2 x 94e SLVER1 PLASTC315 at 2.4 mn 7.3 6375 3.2 6340 ++ 6200 DEFORMA\TION

SPACING (100mmnn)

*Average Peak Value for Accelerometers Located at Panel Centre

**Average Peak Value for Strain Gauges Located at Panel Edges

+ Negligible

++ Accelerometers Not Mounted

c Equivalent in Peak Pressure to a Sinpl- Charge Mass of 454 kg

UNCLASSIFIED

%. % %%

UNCLASSIL ILl) ' 485

gJ.

U,LUi

C,

U.

CL

0J

L N TA SS I -'1'. 1

nc

UNCLASSII.IED SR 485

4.57m GZ

', S26

II I I II

II I I IIII I , ,,

2 2.44 m

I ' I I lv m

III I I

*III I

1- S27 S 29mn 24

* I I I

S28 )II ' ,'- -- - - - - - - - - - -* - - - - - 2 - - - - -

1117 1111111111I I 1 1 1 11 I 1n IR-!

6.35 mm PLATING

76 x 152rmm T BEAMS AT 915 mm SPACING

IiI

,S18 0-S19',S21 0 S22

-II- -- L A7 0 A8=I III

iiii oA oA I

S 515 A2 ''A3 A5 S24I I o

=S14 A1 S160 " 0 A4 A $

II 17 D12 '"13 S S25V

III I I II II I IIII)1 II m 0 A Q 9

I I II I I II I II I I I

I I S I I I

IIII II u IIm

BOTTOM SURFACE

AS St IN I IlHOUG t 1l1

Figure 2

TRANSDUCER LOCATIONS ON STIFFENED PANEL FOR MINOR SCALETRIAL (P PRESSURE, A = ACCELERATION, D DISPLACEMENT,

S STRAIN)

I'N(I \SSI111l1)

"-,,,

UNC[ASSIFIED) SR 485 0

Figure 3a 22

STIFFENED PANEL STRUCTURE BEING INSTALLED ON REINFORCED '

CONCRETE FOUNDATION FOR THE MINOR SCALE TEST U

Figure 3b

PANEL AS INSTALLED FLUSH WITH THE GROUND WITH DRES

BLAST-GAUGE STATIONS AT EITHER END OF THE PANEL

('V I sl~llItU-

% % ?

U N('1A,-S I. I 1.[1) SR 485

MAGNIFICATION =4.00

CENTRAL64 DEFLECTIONS (mm)

NOTE: ARROWS INDICATESTIFFENER POSITIONS

zY

X

Figure 4

PANEL DEFORMATION AFTER FINAL SHOTOF DRES HOB TESTS

(Damage Exaggerated by Expanding the Vertical Scaleby a Factor of 4)

N NN

%>al % %

-o . S N i AD. .. . . . . . .... - ,* . . -:. ;.

t N _ .\SS.I I ) SR 485

S 18, S19 '

09m

37 A75;

S 2029 A127

,14

SI2 15 253 S17TE303~~ 12 A3 A32l1

0.9"15 m °1"

r - - ,T

I I I

II I I I _ _

'N N N N

, a , imiiw

UNCLASSIFIED SR 485

STATIC LOAD ANALYSIS -400- PANEL CENTER DISPLACEMENT

ENTIRE BEAM IS YIELDING

-.. ' (NUMERICALLY UNSTABLE)

300.

0

o -

-j

YIELDING

u 200

0

cc

100-

014.0,

10 50 ;5 10

DISPLACEMENT (mm)

Figure 6

ADINA RESULTS OF DISPLACEMENT VERSUS STATIC LOAD .

FOR STIFFENED PANEL ,UNICLASSIFIE'D

Lu1

',,",." ..." ." - ,, ',"-"."."- 'r"- ," . .-""",h ","", .-"- ,, , ,r,,"",P", ,e"/ % ."" .. ", , - """.. -:" . . ". --.-J.,'_' . , - .

: -" " " |": . , , " f" , t ; I " " "" " m " "'0" ,' -

UNCLA SSIIEDI SR 485

OVERPRESSURE

4.8 KT CASED ANFO CHARGE (DENSITY =.9)

STATION NO. 19

PROBLEM =52.0100

397 kaYMAX= 397 kPa400 -' TIME OF YMAX =0.1962 sec

ARRIVAL TIME = 0. 1958 sec

35kaRADIUS= 278 m

USE MODIFIED CURVE

0.0

w

w0

100

00

0.18 0.24 0.30 0.36 0.42 0.48 0.54 0.60 0.66 0.72 0.78TIME FROM DETONATION (SEC)

0 190 620

TIME AFTER ARRIVAL (MS)

Figure 7

PREDICTED PRESSURE-TIME HISTORY LOADING FOR EVENT MINOR

SCALE AT STATION 19 NEAR SITE OF STIFFENED PANEL STRUCTURE

U NC LASS!IF IED

%................................... . . . .%..

% .

UNCLASSIFIED SR 485

MINOR SCALE TEST

0 NODE :289 S.

E -25E FINAL

z -," , DEFLECTION !-50

a. -75 .

-100 -- T 1-1 1

0.00 0.01 0.02 0.03 0.04 0.05

TIME (SEC)

(a) CENTER OF PANEL (TRANSDUCER r 12);

MINOR SCALE TEST

o NODE 303 k

E

z -25W FINAL

S-, : DEFLECTION-J , ,/ , ,5

0.

Ub-50-

-I I I I :

I

0.00 0.01 0.02 0.03 0.04 0.05

TIME (SEC)

(b) MIDSPAN OF BEAM STIFFENER (TRANJSDUCER D 13)

Figure 8

PRELIMINARY PREDICTIONS OF DISPLACEMENT-TIME HISTORY

UNCLASSIFIED

.... **** ... ~S11 I- • • - I , -l =h

M

'l ' " % % - % " = ' o % ' . % - . . °

L N( LASS I IID SR 4S5

MINOR SCALE TESTNODE 2892000-

Z 1500_

0nn. . . ,£ '

< 100.. .. ,,w-_ .. . : ,..p,,

o

S-500 p

-1000-

r - T T T T T 1 T T T I T f i-1 T T

0.00 0.01 0.02 0.03 0.04 0.05

TIME (SEC)

(a) CENTER OF PANEL (ACCELEROMETER A2)

1500 NODE 303

1000 . .

z AO 500 .

-50Iw , ,'

o 500 - ..

-1000

-1500 "-T T i TT I : t , r t Tir I

0.00 0.01 0.02 0.03 0.04 0.05

TIME (SEC)

(b) MIDSPAN OF BEAM STIFFENER (ACCELEROMETER A3)

Figure 9

PRELIMINARY PREDICTIONS OF ACCELERATION-TIME HISTORY

UNCI ASSIFIEI)

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

'- ,%.% I -... - 5 . %% "--. \.'.r - .".* * . -"r r°.-- p". -". " . - - . . - - - -

a'I

UN(I ASSIFIEI) SR 4s

MAX PRES = 384kPa AT -140 MS

4- 00- CL 400- ' '"

300 Uj 100. P

-5 200 10 2

tY" TTIME((MS

_MAX PRES = 344kPa AT 1.60 MS .4000

I

" 100"

cc 200

-10 -0 510 1520 25 3035 4045

cc TIME ( MS )

0 P21

-100 0 100 300 50 700 900

TIME (MS)

MAX PRES -: 340 kPa AT 2.25 MS

4001 - 400

300

C, 3W 2000 '4c P3

o. -100 ,- o

,t' Or) 5 0 5 101520 2530 3540 45

S0 P3

0.U

-~C 10000."

10100 00 300 500 700 900 ;5,

TIME TMMS ) (,S

MAX PRES = 359kPa AT 1.45 MS400- .5,,. 30000

200 -P

0 " cc 00 -

1 200

LU 200 o~ 0 -""

(1 100 5 0 5 10 15 20 25 30 35 40 45

C{: TIME I MS)- 0.0 0 03

CL

-100 - - -

-100 0 100 300 500 700 900 'v

TIME (MS)

Figure 10

'MEASURED SIDE-ON PRESSURE-TIME HISTORIES ON STIFFENED PANEL s

MSURFACE (TRANSDUCERS P1 - P3) AND BLAST GAUGE STATION

,,: (TRANSDUCER P8) ',C"tN(L \ 1 III)

_Y 3.2

P---------------------------------------------------------------------------------------------------------------------------------'."--.---.---'"-'""V--'-.-8.S100...

;% • q1 '-% - • -"% " "' , ," .cc. . " .,. . , ' 2,00,- ,,..,. . ',."- W, , . . .- - - -, " :

U N(I ASS II-I[-.[) SR 485

zC14

N z

I- w

ww

4. I.- I

% U.)

I- 0

rI-

CDCD

Sza

0

LO 0

CD Q D 0-

w- 00N 00cM- CD u

(P4-.nS 3H

IJN(I AS IFI0

PD r%

UNCLASSIFIED SR 485

MAX DISP = -53.1 MM AT 15.6 MS10

- 10-

S-30 D12

-70 . . . . . . . . .. ..- --200 0 200 600 1 000 1400 1800

ww

~~~TIME ( MS ) ,"

MAX OISP = 29.3 MM AT 20.8 MS v

C-)-0

-20

u 30 ",

-200 0 200 600 1000 1400 1800 .

TIME (MS)

MAX DISP -53.1 MM AT 15.6 MS

10-

10- li IID1

z- D 122. -30

-aa.

_, -50

0 70 . ,J . . .

-200 0 200 600 1000 1400 1800

TIME (MS)

Figure 12

MEASURED DISPLACEMENT -TIME HISTORIES AT PANEL CENTER(D12) AND BEAM MIDSPAN (313)

-30('I,, 1! 1)C-)C

Lsp.

",.,_. rZ, ,-,.;" -50',,v "a-,,'-,--, . -" ._. " . ' " ." "," - ."-" " "," " . . ."---"-"-"-"-"-".",• .• , = __ a .FI = = " * - - " % . ' % ' - " - " . % " % " . % " . " . " %%. % % '

UNCLASSIFIED SR 485 -

a)

I j;Ir

Figure 13

POSTSHO PHOOGRPH O STFFEND PNEL TRUTUR

-. M M

Figure 13b

FINAL DEFPERMANENTF DENELESTRFNS

(Daag EAxIaTIONte byEpnigteVetclSaeb 10)FO

PANELAANDFIEA

-MIDPOINTS

e %I v%, - .. % !..%

UNCLASSIFIED SR 485MAX ACCN 380 G AT 5.60 MS P.

400-

L7 2002 - pj I /N Al1

0 a

• 200WU

4 -400I I T

-10 0 10 20 30 40 50

TIME ( MS)

MAX ACCN = 1310 G AT 14.8 MS -a

( 1000

0

-J

cr A2 .

Uj ,wUJ -1000

I I r I r , , ! ,;

-10 0 10 20 30 40 50TIME (MS)

MAX ACCN 1280 G AT 4.60 MS

a. II( 1000 p-.

zA0

44 0LU aa

-j-LU

U 1000

2000 r , , , --.. ,I-10 o10 20 30 40 50

TIME (MS)

Figure 14

MEASURED ACCELERATION TIME HISTORIES ON STIFFENED PANEL

I N( I \-,,11 11I1)

.....

,' ,% L°. % ' ' % ' ' - ' ° % ' ° ' ' % 'o" .' ' ' % " % ' ' " ° " % ' ' % " % ' ' ° " % ' % . '." .' o°, ".' ' '.' ' % ' % '.' ' % ' . % % %, % = % ", . .% l ' 'a

• w ,P # "" .€ v" " " 4" " " . . € ° " ' . ." . . ." • . .* .' . , " " o= % " ' • w W %3. % .. % • % * '., .. .. % % .,. .. ,,,% -. . -, . -. .,_" % *,,, ,. ,% -.. ,. ',,a,.,,,. %,,,% %, .,, % % % ,,,.',a-% %

-X A. -YJ -J . -. .

UNCLASSIFIED SR 485

MAX ACCN 1940 G AT 3.10 MIS

2000-

~1000- A8 A 7z0

J4 0

U -1000-

-20001 _r

-10 0 10 20 30 40 50TIME (MS)

MAX ACCN =1940 G AT 3.10 MIS2000-

z1000-v/ A

I-I0 p

U -1000

- 200011---r---- _____________

-10 0 10 20 30 40 50TIME (MS)

2000-MAX ACCN 1560 G AT 2.20 MS

1000-

z0

0%

Ui

U-2000- A 6 T

-10 0 10 20 30 40 50TIME (MS)

Figure 15

MEASURED ACCELERATION -TIME HISTORIES ON STIFFENED PANELU NC LASSWI:1ED

e F r

% r

UNCLASSIFIED SR 485

MAX DISP -53.1 MM AT 15.2 MS: 10 1- ,

I- - 10

U A2 P-

-30

S -50 ,. D 125,.

-- 70 F Ti ' v rr, ,y ur . . . . . .r -. . . .. "-2 2 6 10 14 18 22 26 30 34 38 42 46 50TIME (MS)

MAX DISP = -30.7 MM AT 20.4 MS

0

-10 4' /*3 -lO- 3

0100(L -30- 3

- -40!

-2 2 6 10 14 18 22 26 30 34 38 42 46 50 -

TIME (MS)

MAX DISP -50.1MM AT 19.7 MS

S-10 .-

-301 - -" A5LU

' .-U - - ~A4( 501

70 r -VI.r-- ,-

-2 2 6 10 14 18 22 26 30 34 38 42 46 50 1TIME (MS)

Figure 16

MEASURED DISPLACEMENT-TIME HISTORIES FROM TRANSDUCERSD12 AND D13, AND DOUBLE INTEGRATION OF ACCELEROMETER

SIGNALS A2 - A5

I 'N( I \ w;11 I I)

N... ,

Ll N('IA,,SSIII lED SR 485

FW MAX STIR r3 6 6 0 k AT 15.1 MS

0

3

22

0 40 102003060

TIME (MS)

MAX STR 270 AT .0 MS

x 2

cm v

* ~ 0

22

0 4010020030040 00s

TIME (MIS)

M EAUE AX-T HISTRIE AT O 3.70I MSGEoNTPADBTO UFCSO AE

x S._,XS l~F

. . . .. . .

W r e. '- _-:~pz

UNCLASSIFIED SR 485

4-p

2-.

z

S290

0 40 100 2003040TIME (MS)

MAX STR = 2350 yz AT 4.20 MSo 1

- - 2

0- S14

zrtof'

0 40 100 200 300 400TIME (MS)

MAX STR -1960 p AT 1.60OMS

11

* 0-i

* 0 1~I

S-2

0 40 1 00 200 300 400

TIME (MS)

Figure 18

MEASURED STRAIN-TIME HISTORIES FROM STRAIN GAUGESAT VARIOUS LOCATIONS ON THE PANEL

UN('I.SSII 1.1

% %~

0IUNCLASSIFIED SR 485

MAX STR -1960 tw AT 1.60 MS

0

Io~2 40 100 200 300 40

MAX STR =-l9 OML_ AT 20.3 MS

0 S23t

I ~ S~ ~a~v~afS22a. I

* 0oJ -2-

0 40 100 200 300 400

TIME (MS)

C.'MAX STR = 1960 pi AT 1.60 MS

0

It IS 20,

S 18

0

S-2

0 4 1 0 20 30 40 50

TIME (MS)

Figure 19

MEASURED STRAIN-TIME HISTORIES FROM STRAIN GAUGESON FRONT AND BACK EDGES OF THE PANEL

UNCLASSIFIED

*~~~ Jr N* -.. - - - - - - - - - - -

UNCLASSIH'ED SR 485

A7 (8)

z (16)

(457

4' D12,A2

(46

(472

D13, A3, (Al1)

4'4' FINITE ELEMENT MODEL

I-8

457 146

UNC ISI ij--1 )I I %

U NCLASS III L. 1 S R 485

MINOR SCALE -DETAILED FINITE ELEMENT MODEL 11-09-85

0p

I -25-wj -

C.)

U~-50-

0.0 0.01 0.02 0.03 0.04 0.05

TIME (SEC)

(a) CENTER OF PANEL (TRANSDUCER D12)

MINOR SCALE -DETAILED FINITE ELEMENT MODEL 11-09-85

0~-

-25.

z

C-ch -50

0.00 0.01 0.02 0.03 0.04 0.05

TIME (SEC)

Wb MIDSPAN OF BEAM STIFFENER (TRANSDUCER D13)

Figure 21

FINAL PREDICTIONS OF DISPLACEMENT-TIME HISTORYUNCLASSIFIED

% C%.rdC ~ C C d~ - .'- ..- .- . : . r .t V -

7 717% W 70 7VI

UNCL.ASSIFIED SR 485 %

MINOR SCALE - DETAILED FINITE ELEMENT MODEL 11-09-85 %

NODE 457 I

1500

- 1000,0

500.."" - : . . :' . : " " "

LU 0.

-500

-1000 -_,-_I I I I I I I I I I I I I

0.00 0.01 0.02 0.03 0.04 0.05

TIME (SEC) "

(a) CENTER OF PANEL (ACCELEROMETER A2

MINOR SCALE - DETAILED FINITE ELEMENT MODEL 11-09-85-. '

NODE 472

."1500I- '..

'4."

0 , ,,%,

w --- I500 4,

4 - 500 v5

-1000 :I

000 0.01 0.02 0.03 0.04 0.05'.=

TIME (SEC)

W(b MIDSPAN OF BEAM STIFFENER (ACCELEROMETER A3)

Figure 22

FINAL PREDICTIONS OF ACCELERATION-TIME HISTORY

UNCI ASSIFII-I)

4'._

".. ... ' , 4.- *. . 5# . ".. - a -....°. ..% . ..

% -- .- -

tN~jI:\SSIJLI)SR 485

MAX ACGN -1460 G AT 7.50 MS

0 1500-*'

0' A 2

i I

U ~A 2(FE)

-1500, I I I I I IT~. I I T

- 4 04 8 12 16 20 24 28 32 36 40 44 -

TIME (MS)

MAX V EL = 18.1 MIS AT 14.1 M S /

10-I

0- A 2 (FE

I --- *

>A_10- A 2

- 4 0 4 8 12 16 20 24 28 32 36 40 44

TIME (MS)

10 MAX DISP =-53.1 MM AT 15.2 MS

-302A

-s - D 12-5

- 4 0 4 8 12 16 20 24 28 32 36 40 44 A

TIME (MS) A

Figure 23

COMPARISON OF FINITE ELEMENT PREDICTIONS WITH THE MEASUREDAND INTEGRATED MOTION-TIME HISTORY AT CENTER OF PANEL

(TRANSDUCERS A2 AND D12)

UNCIA,,SSlI ILL)

Al V lI-~.IA. A'

UNCLASSIFIED) SR 485

MAX ACCN =795 G AT 6.00 MS

800

0 A0

MA-8008.4MS T5.0M

4

A3A

* 00

-4-

-4 0 4 8 12 16 20 24 28 32 36 40 44TIME (MS)

MAX DISP 30.7 MM AT 20.4 MS

0.

w-10A 3(FE)

Q -20

~ 30- A3

04

- 4 0 4 8 12 16 20 24 28 32 36 40 44 -

TIME (MS)

Figure 24

COMPARISON OF FINITE ELEMENT PREDICTIONS WITH THE MEASUREDAND INTEGRATED MOTION-TIME HISTORY AT MIDSPAN OF BEAM

STIFFENER (TRANSDUCERS A3 AND D13)

% r .6 e

r* del p- - .5.-'.

L'NCLAS'SII III) SR 485

-S20 [S18] (NODEl)S22 S2FIE

S23 BOUNDARY

(6)

z (10)

S21.

(133)

D 13, A 3.[l2*EUVAETLCTO

S1I.

S15.

(142)

UNCLASSIFIED SR 485

z

z0

-

w%

o Z "U

0U-

a,z

0

J

LLa

o -C.l

t80

S---0.

w 0 0 0 0 00 0 0 0 0a

(2d)l) 3EJnSS38d

UNCLASSIFI IlD

UNCI-ASSII II) SR 485

0 EXP .

-10 i /

-20-

,, -30<

0.A

S-40- FEFE

-50-

-1 -- A 7 ' r '' 1

0 100 200 300 400 500 600 700 800 900

TIME (MS)

(a) CENTER OF PANEL (TRANSDUCER D12)

I1.

-4-

EXP

- E/FE

_' -24 - FE

(:O-208'

-32

0 100 200 300 400 500 600 700 800 900

TIME (MS)

(b MIDSPAN OF BEAM STIFFENER (TRANSDUCER D13) -

".-

Figure 27

COMPARISON OF FINITE ELEMENT ADINA RESULTS WITHMEASURED DISPLACMENT-TIME HISTORY

UNCI.ASSIIEIDI)

. ... . . . . . ..,Y

,..w ., -.., .- .' .-_. .,= - ',, ,. ., ,. .', *5 .. . . . . -, -.".." . . . 5. .2... . ." - .' " - -"- - -- - - - . -: - ." ,., "

UNCLASSIFIED SR 485MAX STR =1860 pz AT 26.7 MS

2

-2 2 6 10 14 18 22 26 i0'34 8 42 46'50

TIME (MS)

MAX STR -6530 p AT 7.00 MScnS

Q 6 S\ // S27 (FEA)

5%

c0cc 0

-2

-2 2 6 10 14 18 22 26 30 34 3 2 45

TIME (MS)%

4 MAX STR = 988 jAE AT 14 MS

o-

0

S15 (FEA)

-12

-2 2 6 10 14 18 22 26 30 34 38 42 46TIME (MS)

Figure 28

COMPARISON OF FINITE ELEMENT ADINA RESULTS WITH MEASUREDSTRAIN-TIME HISTORIES ON PANEL NEAR BEAM MIDSPAN

UNCI \SSIF 11I:1)

%N

% J. e40 %

UN(L. ,SSI1J: 1 1) SR 45.?.:

MAX STR -1960ju AT 1.60 MS

0S1

o S18 (FEA),

- 2 2 6 1 14 18 22- 26 3,0 34 38 42 46 50 .

TIME ( MS )

4 -MAX STR -365 pE AT 3.00 MS

1'

2 -2

-- S19 (FEA)- 1 I I

-2 2 6 10 14 18 22 26 30 34 38 42 46 50

TIME (MS)

MAX STR = 2990 AT 16.5 MS

3 -

,-~~ /.S26 (FEA)

- S 26

0 /

r I I I I . I

2 2 6 10 14 18 22 26 30 34 38 42 46 50

TIME (MS)

Figure 29

COMPARISON OF FINITE ELEMENT ADINA RESULTS WITH MEASURED

STRAIN-TIME HISTORIES ON PANEL NEAR CLAMPED BOUNDARY

UL(.ASSIFIEI)

C~ 3

= . a . a s . .. . . - . _ - q.--a

UNCLASSIFIED SR 485

MAX STR = -1670, AT 40 MS

0 S20C., 0

\ i . i

cc

_o -2~,p.

-2 2 6 10 14 18 22 26 30 34 38 42 46 50

TIME ( MS)

MAX STR =2990 pE AT 16.5 MS

-. S 26 (FEA)

2

cc.

o 0U I

-2 2 10 14 18 22 26 30 34 38 42 46 50

TIME ( MS)

:41

MAX STR -1810/4E AT 9.20 MS

0'" S 18 ( FEA )

0 S.23

2

-2 2 6 10 14 18 22 26 30 34 38 42 46 50

TIME (MS)

Figure 30

COMPARISON OF FINITE ELEMENT ADINA RESULTS WITH MEASUREDSTRAIN-TIME HISTORIES ON PANEL NEAR CLAMPED BOUNDARY

EQUIVALENT TO GAUGE LOCATIONS S 18 AND S26

UN(I ,\SSI I1I)

L'N(I\SS~IIL)SR 485

FIXED NODElEND

2

3

4 -

5

1219 mmn6

7

8

CENTER

- 76K

Figure 31 a/

FINITE ELEMENT EQUIVALENT BEAM MODEL OF 2286 mmSTIFFENED PANEL FOR FENTAB

,IL

2438 m

Figure 31 b

FINITE STRIP MODEL OF STIFFENED PANEL

U NCL ASSII FID

., j* **

LJNCI.,\SS I !I ED SR 4850

-I" I" " - 10 PANEL

Z )0 ADINA

-FS--30 r!

S0 40 / "

50 EXP

'5I II I0 20 30 40 50

TIME (MS)

(a) PANEL CENTER (TRANSDUCER D12) USING FINITE STRIP METHOD

-- 4-

BEAM-8ADINA

Z -12j -16 FS

, . . , . .. .\. ...

0-24-

-32- T-----0 10 20 30 40 50TIME (MS)

(b) BEAM MIDSPAN (TRANSDUCER D 13) USING FINITE STRIP METHOD

0-

-,. BEAM

- -.,

-12- ADINAV z

"" -16 I ,

-20" FENTAB

0'; ",\ ... ;;)? .. J.< 20

-28 EXP

-32

0 10 20 30 40 50TIME ( MS)

(c) BEAM MIDSPAN (TRANSDUCER D13) USING FINITE ELEMENT

EQUIVALENT BEAM METHOD FENTAB

Figure 32

COMPARISON OF UBC NUMERICAL PREDICTIONS WITH ADINARESULTS AND MEASURED DISPLACEMENT-TIME HISTORY

.5

:_ P ';"',', ' "e,"/"". .- "-,.- . "w,'- , ". ., *%,.-.- -, --.- .* . .* * '

U NCiLA\SS I IILD SR 48S

F S-10- X

S-20

(L ADINA40

-50 1

0 100o 200 300 400

TIME (MS)

(a) PANEL CENTER (TRANSDUCER D 12)

0-

FS-10 X

CL

-30

0100 200 300 40

TIME (MS)

(b) BEAM MIDSPAN (TRANSDUCER D 13)

Figure 33

COMPARISON OF THE FINITE STRIP PREDICTIONS WITH ADINARESULTS AND MEASURED DISPLACEMENT -TIME HISTORY

UNCLASSIH-EL)

°..

B - 4 [INCIASSTFIFID

This Sheet Security Classification

DOCUMENT CONTROL DATA - R & 0ISecu,,Iv .1es.C*olion of title, body of abstract and Indexing annotation must be entered when the overall document is clawseiadl

ORIGINATING ACTIVITY 2s. DOCUMENT SECURITY CLASSIFICATION

Defence Research Establishment Suffield _nla-,_ifi_ JedRalston, Alberta TOJ 2NO 2. GROUP

.I DOCUMENT TITLE

STRUCTURAL RESPONSE OF SHIP STIFFENED PANEL AT EVENT MINOR SCALE (U)

4 DESCRIPTIVE NOTES IType of report and inclusive dateOlSuffield Report No. 485 (1984-1987) r:.

AUTORISI ILest name. first name. middle InitIal)p, .

Slater, J.E., Houlston, R. and Ritzel, D.V. p..'

6 DOCUMENT DATE 7s. TOTAL NO. OF PAGES b. NO. OF REFSJanuary 1988 65 J 38

&L PROJECT OR GRANT NO. go. ORIGINATOR'S DOCUMENT NUMBERIS)

0311R SR 485

Bb CONTRACT NO. 9b. OTHER DOCUMENT NO.IS) (Any other numbers that may beassigned this document)

10 DISTRIBUTION STATEMENT

UNLIMITED I

II SUPPLEMENTARY NOTES j 12. SPONSORING ACTIVITYTask for Director Maritime Engineering Suppo t i.

National Defence Headquarters, Ottawa

13 ABSTRACTModern computational methods and experimental test procedures

are being developed to determine more accurately the air-blast loadingand dynamic response of naval ship stcuctures subjected to weapon %explosions. Previously, a series of air-blast trials on full-scalestiffened panels typical of those used in warship superstructures hadbeen conducted for conventional scale weapons (under 454 kg HE) toevaluate new design and damage-assessment techniques. The presentreport describes the experimental results and the associated numericalstudy of the panel response for a longer duration loading generated by N.

the simulation of a 8 kiloton nuclear weapon surface explosion calledMINOR SCALE. The panel, exposed to a side-on overpressure blast load-ing of 345 kPa peak and 200 msec positive duration, experienced onlyslight permanent deformation of about 25 mm with no apparent weld orbuckling failure. The measured pressure, strain, acceleration and dis-placement data correlated well with the computed numerical results,thus verifying the computational methods.

~I,. .A M

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

UNCLASSIFIED_______________This Sheet Security Classification

KEY WORDS

Structural ResponseAir-BlastPanel (s)

?v.-

a.

INSTRCTION

('Hl INIIN, A' TVITYEntr te nme nd adres o th qb OTER DCUMNT UMBR(S It he ocuenthasbee

,,e~-q ho d.:unen assgne en othr dcumet nmbes fathe bythe or~ato

or b thesposor) alo ener tis umbe~s)

trermiIA vyoeeoi to-aCIITY Enfuhrnm n adesof~% ther dOCsemntno NthMeR Iouet. ther dtChmnt hs be

by security classification. using sandard statements such as.21i, ('.RIIUP F iti 5i'ruiity reclassificationi group number. The three

tjijii, iii iiit-odi~ it Appendir Moll th" ORB Sescurityo Regulations. fil Ousliied requesters may obtain copies of thisdocument from their defence documentation center."I

I f)I)ClJMFNT II iILLEmIrile the c~omplete document titls in all00a ruth- miri It' ill caries shorid be unclassified 1f a 12) "Announcement and dissemination of thq document

5itdioi n fic..pt-e leter cannot be selected without casif- is not authorized without prior approval fromli s t-w tlt. , issificatiofl with the usual one copotel-llet originating activity.'

Ii i.r iiipu iiithoses immediatrriy following the title.11. SUPPLEMENTARY NOTES Use for additional exlanatory

4 )1 ,(-'RIPT yVE NOTIS Eliter the Catiegory of document. e.g. notes.-to 'n..oo rerfehniscri note or technical letter If appropriit,-,t thi tlp. I ioruinent. e.g. interim, progress. 12. SPONSORING ACTIVITY. Enter the name ot the departmental

n.. nis ifiniusi , finial Givoo the inclusive dates when a project otfice or laboratory sponsoring the research andolo. rportig Period ii covered, development. Include addreas.

5 AUTH(1IIS Eir,i the. iurnalsl of outhorltl as shown on or 131 ABSTRACT Enter an abstract giving a brief antI factual, hr nrititirneit Eoe Io tis name. first name, iddle initial summary of the document, even though it may also appear

It -- tIw show ril IThee narme of the principal author it en. elsewhere in the body of the docurnent itself It is highly,ili-Aoitri minrmum,, isiji ,ir,nont desirable that thrr abstract of classvifloed documents be unclass-

toed. Etich paragraph of the abstract shall end with an6 DO(uMENT DATE Firer, the fare0 (month, year) of indication of the security classification of the information

F %isrt~l.%hirrni finoai for publication of the document. in the paragraph (unless the document itself io unctassified)represented as ITS). IS). (Cl. (RI. or (U)

T 0TA L NUMBE R OF PAGES The total page count shouldnrii iiinnl pegrirrrrcircedures. i ao. enter the number The length ot the abstract should be limited to 20 sinigle-spaed

,I Iingie Ii finig rl.cinaocn standard typawrelirinrtin lines 71,% inches long.

it 'JtiMbiI or I Il ifPL It[ NI ES Ente the toital numribar of 14. KELY WORDS Key worrfs are technically meaningful terms or,,tri,,,,irr ,it' I.. -- ofrr short phrases that chirrterire n document eind voirlrt he hislpful I'

54~ ii~ F I li tntAN H t eterthein cataloging the document Key words should bit seler tadtsoMe '4j~iItl ,ININUMBENI altxpriatsenorth thast rio erecur ity classiticietion is required Idenifiers. %uch as

lupii ahl.. -- too-, 1 iiweliounste pboiect or grat number Pilurpnrirrtt mortal dlesignatrion. ririle rame, mrlifsry poolect corde-I.. 1,f, If-i., l-. -m-1 i viwe, wrilteri conete. geographic location. nay tie used oes key worth but will

lie followelod by an irdicaution of trichnoicjal coite. tMob ON Ti A( I Nklf MII I , jjp~npnis ii owrie the Pipplirablie

-'I.. . , s.,h hr. lio-imernt wirs written

on t)AIGINATOII S u(JCUMI-NT NUMBEHIISI Ente thel*0-oo tn f..rirt. inor,i -Iii......tiri, try which th. document will tIe

.,,I i nw-Iitlpr Iry ihe origrirating activity This e e'ier r'st tie itrsuoint this docirment

I fe I r I ..

%~~~~a Ne e.P % N%%N

- - - -°- "

i .!v

*

*, *

-sm J

.,....