ML20126F467
| ML20126F467 | |
| Person / Time | |
|---|---|
| Issue date: | 12/21/1992 |
| From: | Taylor J NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO) |
| To: | |
| References | |
| SECY-92-424, NUDOCS 9212300327 | |
| Download: ML20126F467 (70) | |
Text
[
O O
O O
O 0
.M M
n n
n n
M M
M R
O e 0 e e e t e 0 6
- e e o e e 4 o O G O O O 4, RELEASED TO THE PDR O
,e 6 2/93 6:
O r
N we' we g
1 0o00000000o000000000000o 1
\\,,,,,/
i POLICY ISSUE December 21, 1992 (NEGATIVE CONSENT)
SECY-92-424 f_qt:
The Commissioners From:
James M. Taylor Executive Director for Operations Subiect:
PHASE ONE EXTENSION FINAL REPORT, " FEASIBILITY STUDIES FOR THE CONTROLLED, CRUISE-ALTITUDE DROP, AND AIRCRAFT CRASH TES1S" Puroose:
To inform the Commission of the staff's intent to forward the Phase One Extension Final Report to the Power Reactor and Nuclear fuel Development Corporation (PNC) of Japan.
Backaround:
The Murkowski Amendment (Sec. 5062, Public Law 100-203) imposed requirements on the packages used for the air transport of plutonium from one foreign country to another, where such transportation passes through U.S. airspace.
Previous Commission papers on this subject have addressed:
the Amendment's requirements (SECY-88-302); the Agreement between the Nuclear Regulatory Commission and PNC on the Testing and Administrative Program for Certifying Packaaes (SECY-88-329); the selection of a worst-case aircraft crash s
(PSA Flight 1771) for test criteria development pur)oses (SECY-88-344); draft package drop and aircraft crasi test criteria (SECY-89-208); interim progress (SECY-89-382); and submission of Phase One Final Report (SECY-90-301).
The Murkowski Amendment requires that a plutonium air transport (PAT) package undergo the following tests without rupturing or releasing its cor. tents: (1) an actual drop test
Contact:
NOTE:
TO BE MADE PUBLICLY AVAILABLE WHEN THE FINAL SRM IS MADE 4
AVAILABLE I
i bH d 1 x 1 1 1 s s s s s s ). R C A ogyg 8'n n
\\
t The Commissioners 2
and (2) either an actual crash test of a cargo aircraft; fully loaded with PAT packages, or "other" tests that the Commission dc' ermines, after consultation with an independent scientific panel, exceed the stresses that would-occur during a worst-case plutonium air shipment accident.
In Phase One of the Agreement between NRC and PNC, draft criteria were developed for conducting tests to satisfy these requirements.
Phase One was extended to perform feasibility studies for conducting these tests.
l As specified in the Agreement between NRC and 'NC, a final
(
briefing on Phase One Extension activities wat irovided to
/
PNC on September 16, 1992, at PNC' headquarters in Tokyo, 7
Japan.
Q11cug11gn:
The test criteria for the "other," or controlled, test, developed under Phase One, are based on in6 worst-case aircraft accident on record in commercial r.viation, PSA Flight 1771. The test criteria require a perpendicular impact of a package onto a target surface hardness of 2.5 (i.e., soft rock) at e velocity of 282 m/s (925 ft/s).
Under the extension, the contractor, Lawrence Livermore National Laboratory (LLNL), investigated the feasibility of performing a controlled test, at a lower velocity, onto an essentially unyleiding surface and obtaining a pack 2ge response that is equivalent to, or greater than, the maximum G 1evels it would sustain from an impact test of the prescribed conditions..LLNL performed computer analysis and benchmark testing to evaluate this approach.
Also, under the Extension, LLNL identified facilities and organizations available for performing a controlled test to meet the prescribed conditions, as well as facilities for conducting a cruise-altitude drop test.
LLNL also identified a method for conducting a single test in such a manner that the test criteria specified for the cruise-altitude drop test as well as the. controlled test could be satisfied.
Performing a single test to satisfy the requirements of the controlled and the cruise-altitude drop tests'would be subject to review by an independent c ent fic panel.
lhi, report concludes the technical efforts specified in Phase One of the Agreement between NRC and PNC.
There have been no indications by PNC that it intends to continue 1this activity.
f l
N 4
The Commissioners 3
Schedulina:
PNC has requested that this report be forwarded to it on December 31, 1992. A written estimate of expenditures of Phase One and Phase One Extension activities will be provided to PNC by March 31, 1993, followed by a final written explanation of expenditures by September 30, 1993, Coordination:
This paper has been coordinated with the Office of the General Counsel, and it has no legal objection.
Recommendatipn:
In accordance with the schedule of the NRC/PNC Agreement, the staff intends to provide PNC with a copy of the Final Report on December 31, 1992, unless the Commission directs otherwise.
es M. T -lor i
xecutive irector for Operat!ons
Enclosure:
(Commissioners,SECY, and OGC only)
Phase One Extension Final Report -
" Feasibility Studies for the Controlled, Cruise-Altitude, and Aircraft Crash Tests" i
SECY NOTE:
In the absence of instructions to the contrary, SECY will notify the staff on Tuesday, December 29, 1992, that the Commission, by negative consent, assents to the action proposed in this paper.
.'STRIBUTION:
< :nmissioners GC OCAA OIG IP OPP EDO ACRS SEC'
i' i
UCRL-ID-112582 i
6 i
Plutonium Air Transport Certification (PATC) Program Phase One Extension Final Report B
Feasibility Studies for the Controlled, i
Cruise-Altitude Drop,.
i l
and Aircraft Crash Tests J.H. VanSant, T.F. Chen, and LE. Fischer I
i j
Prepared for:
U.S. Nuclear Regulatory Commission
(
l lli LAWRENCE LIVERMORE NATIONAL LABORATORY l
. - ~.
1 i
)
DISCLAIMER l
This document was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Govenunent nor any agency thereof, not any of their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any iniormation, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any
?
specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, l
does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof, t
i i;
f i
t l
4 I
d M
N i
i l
1 i
d I
4 This work was supported by the United States Nuclear Regulatory Commission under a Memorandum of i
Understanding with the United States Department of Energy.
t 1
= -.
4 i
UCRL-ID-112582 i
i l
Plutonium Air Transport Certification--(PATC) Program Phase One Extension Final Report t
i Feasibility Studies. for the Controlled, Cruise-Altitude Drop,-
and Aircraft Crash Tests i
Manuscript completed: December,1992 Prepared by:
1 J.H. VanSant, T.F. Chen, and L.E. Fischer Lawrence Livermore National Laboratory 7000 East Avenue l-Livermore, CA 94550 Prepared for:
4 Division of Safeguards and Transportation Office of Nuclear Material Safety and Safeguards U.S. Nuclear Regulatory Commission
)
Washington, DC 20555 NRC FIN No. L1056 4
--,,,,-,a
~,
,-,.- ~
--.p.
I' ABSTRACT t
4 j-Section 5062 of Public Law 100-203 states that the Nuclear Regulatory Commission -
(NRC) shall certify to Congress the safety of containers used in the air transport of l
plutonium through U.S. airspace from a foreign nation to_a foreign nation by.
L requiring testing as follows: (1) an actual drop test from a maximum cruising altitude of a full-scale sample of such container with test materials, and (2) an actual crash test of a cargo aircraft fully loaded with full-scale samples of such container i
loaded with test materials unless the Commission determines, after consultation -
with an independent scientific review panel, that stresses on the container _ produced
[
by other tests used in developing the container exceed the stresses that occur during a worst-case plutonium air shipment accident. Under Phase One of the NRC l
program for Public Law 100-203, draft criteria for performing the specified tests _were developed and published in References 1 and 2 of this report. Phase One was -
extended to perform feasibility studies for conducting these tests. This report j-describes the results of the Phase One Extension activities. Facilities and organ zat ons capable of conducting the tests are identified. Performance of a single i
i
~
l test to satisfy the required cruise-altitude test and worst-case accident is' discussed.
Finally, the feasibility of conducting tests at a lower velocity on an unyielding j
surface rather than soft rock is explained.
t l
4 i
l r
l l
l i
l l-l
- ili -
_... _, u.. ~ c... ;
EXECUTIVE
SUMMARY
Section 5%2 of Public Law 100-203 states that the Nuclear Regulatory Commission (NRC) shall certify to Congress the safety of containers used to air transport pluto-nium through U.S. airspace from a foreign nation to a foreign nation. Specifically, the law states that the NRC shall require testing with full-scale samples of the plu-tonium container: (1) a drop test from the maximum cruising altitude of the cargo aircraft and (2) a crash test of a cargo aircraft fully loaded with sample containers, or other tests (also named Controlled Tests) that develop stresses in the container greater than would occur during the aircraft crash test. These requirements are in addition to those specified in Public Law 94-79.
In Phase One of the NRC program for certifying plutonium air transport (PAT) packages in accordance with Public Law 100-203, draft criteria for the specified tests (Cruise-Altitude Package Drop Test, Aircraft Crash Test, and Controlled Test) were developed by LLNL for the NRC. At the conclusion of this effort, Phase One was ex-tended to perform feasibility studies for conducting the tests (a task defined as Phase One Extension). This report describes the results of the Phase One Extension studies.
The principal task in the Phase One Extension program is Controlled Test feasibility studies. Included are computer analyses of model PAT packages and physical benchmark tests with package models. The primary objectives of the studies are to validate the methodology developed in Phase One to determine impact velocities on yielding and unyielding materials for equivalent package response (i.e.,282 m/s impact on soft rock surface or equivalent velocity on unyielding surface). Also, two methods for conducting Controlled Tests to certify actual PAT packages are re-viewed. Other tasks include a feasibility assessment of a single test satisfying both the Cruise-Altitude Package Drop Test and the Controlled Test requirements, and a status review of three technical issues pertaining to the Aircraft Crash Test: (1) al-ternative test aircraft, (2) flutter avoidance, and (3) flight control.
In the Controlled Test studies, only one energy absorbing material is considered.
This material is grout, which is a specified mixture of sand and cement. Results of the studies indicate that a PAT container subjected to impacts on a yielding and an unyielding surface can experience equivalent maximum deceleration values.
However, package deformation patterns resulting from impacts on two surfaces would not be equivalent due to interaction with impact surfaces of different proper-ties. The test model used in the benchmark tests did not rupture when subjected to 282 m/s impacts on simulated soft rock and '157 m/s on steel plate. Benchmark test results were incorporated into material property models to improve the accuracy of analytical predictions of package response to impacts. Predicted decelerations and deformations nearly agree with test results, but additional improvements are needed, especially for impacts on yielding targets.
Review of rocket-sled and air-drop test rnethods reveals that these methods can fea-sibly be used to conduct PAT package Controlled Tests. Finally,it appears that the required Cruise-Altitude Drop Test can be conducted such that it could also be demonstrated to an independent review panel that stresses in a plutonium con-tainer would be greater than would occur in an actual Aircraft Crash Test.
- IV -
J i
- 1 ACKNOWLEDGMENTS The authors wish to acknowledge technical contributions to Controlled Test studies 4
l described in this document by J.C. Chen of the Lawrence Livermore National.
Laboratory and W. Hsu of Kaiser Engineering. Also, acknowledgment is extended to j
G.T. Carpluk, M.J. Martini, F.E. Sator, and D.L Turpin_ for contributions during the l
benchmark test program and to B-Division staff for use of the 155 mm gun test j
facility. The authors also wish to thank C. MacDonald and M. Lusardi of the j
U.S. Nuclear Regulatory Commission for their support and comments during the research program. Appreciation is extended to L. Bogart, M. Carter, S. Murray, and -
i l
S. Wilson for document preparation and D. Sceales for editing.
k i
l
(
{
4 e
e i
k l
r I
l
-v-l a
r
---..-n
,;.,i J.,m..
,,'m....c.
__u,_,_,w.-_,,,.w...%
.3,.
--e.-.,
y y
1,
1 4
i a
TABLE OF CONTENTS I
A B STRA CT.............................................................. i i i -
EXECUTIV E S UM M A RY................................................. i v 4
i A C KN O W I.E D G M E NTS................................................... v e,
4 i
- 1. INT R O D U CTI O N....................................................... 1 1
i 1.1 B a ckgro un d......................................................
1.2 Ph ase One Ext ension............................................... : 1 4
- 2. CO NTR OLL ED TEST....................................................
2 :
l 2.1 - Compu ta tional Analyses...........................................
3 -
i 2.1.1 Me thod ology.................................. -............
3 2.1.2 Compu t er M odels..... -...................................
5 i
2.1.3 Simula ted Impacts.........................................
8 2.1.4 Model Design for Benchmark Tests.......................... 8 2.2 Benchmark Tes ts................................................
10 2.2.1 Grout Property Measurements.............................
10 2.2.2 Impa ct Tes ts............................................... 10 l
2.2.3 Impact Test Results.......................................
13 2.3 Analysis of Test Results..........................................
14-2.3.1 Correlation with Analytical Model.........................
14 j
2.3.2 Development of Equivalent Impact Conditions............... 15 i
2.3.3 Correlation of Analytical Model Using Test Results.......... 15 j.
2.3.4 Equivalent Impact Condition Generated by Analysis.........
16 l
2.4 Assessment of Controlled Test Studies.............................
16 i
2.5 Dis cussi on......................................................
17
- 3. CONTRO LLED TEST METH ODS........................................ 33 -
[
3.1 Rocke t Sl e d..................................................... 33 3.2 Aircra f t Drop.................................................... 33 I
i
- 4. CRUISE-ALTITUDE PACKAGE DROP TEST.............................. 36 i
i 4.1 Cruise-Altitude Drop and Controlled Test Impact Conditions........
36 ~
4.2 Test Range Evaluation............................................ 36 -
- 5. AIR CRA FT CRAS H TEST............................................... 38 5.1 Alternative Test Aircraft.......................................... 38 -
5.2 Fl u tter A void an ce............................................... 39 5.3 Flight Control Review............................................ 39 -
- 6. R EFER EN C ES......................................................... 41 i-
. vi.
t' k
r
~,,., _,,,, -.. - -.. -,
l APP ENDIX A S CA LIN G L AWS.......................................... 44 APPENDIX B G R O UT D ESCRI PTI O N....................................
48 APPENDIX C GROUT PROPERTY MEASUREMENTS.....................
50 APPENDIX D OTHER ENERGY-ABSORBING M ATERI ALS................
53
)
l 1
- vii -
L INTRODUCTION The purpose of this report is to document the results of studies performed to assess in specific detail the feasibility of conducting tests in accordance with the criteria given in Refs.1 and 2 (see Section 6, References). The criteria were developed for i
the tests specified in Section 5062(b)(2) of Public Law 100-203 (Transportation of Plutonium by Aircraft Through United States Airspace, Ref. 3).
1.1 Background
j Public Law 100-203 [Ref. 3] states that the Nuclear Regulatory Commission (NRC) shall certify to Congress the safety of containers used in the air transport of plutonium through U.S. airspace from a foreign nation to a foreign nation.
Specifically, the law states that the NRC shall require the following:
(a) an actual drop test from maximum cruising altitude of a full-scale sample of such container loaded with test materials; and (b) an actual crash test of a cargo aircraft fully loaded with full-scale samples of such container loaded with test material unless the Commission determines, after consultation with an independent scientific review panel, that stresses on the container produced by other tests used in developing the container exceed the stresses that occur during a worst-case plutonium air shipment accident.
These requirements are in addition to those specified in Public Law 94-79 [Ref. 41 and NUREG-0360 [Ref. 5].
In Phase One of the Agreement between the NRC and the Power Reactor and Nuclear Development Corporation (PNC) on the Testing and Administrative Program for Certifying Plutonium Air Transport (PAT) Packages, draft criteria were developed for the tests specified in Public Law 100-203. These criteria are: (1) Cruise-Altitude Package Drop Test [Ref.1],(2) Aircraft Crash Test [Ref.1], and (3) Controlled Tests [Ref. 2]. (The "other tests"in paragraph (b) above are called Controlled Tests.)
As part of Phase One, numerous evaluation studies were also performed for development of the test criteria [Refs. 6 & 7].
1.2 Phase One Extension Phase One of the Agreement was extended to perform technical feasibility studies for conducting the tests required by Public Law 100-203. The scope of the studies is divided into four main tasks: (1) Controlled Tests, (2) Controlled Test methods, (3)
Cruise-Altitude Package Drop Test, and (4) Cargo-Aircraft Crash Test.
The first task includes computational studies of model package impacts on various surfaces to develop methodologies for establishing equivalent impact conditions.
Also included are physicalimpact tests with package test models to benchmark the.
F 1
computational results. The chosen energy absorbing-material in package models for h~
j-the Controlled Tests studies is grout because its mechanical properties are isotropic L
and property data are available. Only end impacts of package models are studied.
The second' task reviews suitable methods and facilities for conducting Controlled Tests. LThe third task assesses the feasibility of a single test satisfying both the Cruise-Altitude Package Drop Test and the Controlled Test requirements. The fourth task-develops general guidelines for selecting an alternative test aircraft and reviews the technical status of aircraft flutter avoidance and remote flight control systems.
t
- 2. CONTROLLED TEST I
As stated in Ref. 3, if approved by the NRC for a' specific PAT packaging design, a Controlled Test may be performed in lieu of an ' Aircraft Crash Test in the l
certification process. This option allows an applicant to select an alternative avenue l
of package testing for certification.
The principal elements of the Controlled Test criteria are outlined in Ref. 2. They 1
are:
I impact: The test package shallimpact approximately perpendicular onto an effectively flat target at a velocity not less than 282 m/s (925 ft/s). Package impact i
orientation (e.g., end, side, corner) shall be the one that results in maximum damage -
l to the container at the conclusion of the impact test.
i Tarcet hardness: The effective hardness of the impact target shall not be less than l
that of the identified worst-case impact accident site.
i
[
Eauivalent impact: An applicant may_ conduct the impact test defined above at a L
lower impact velocity onto an effectively unyielding target. Should this option be
[
chosen, the application shall~ determine the lower impact velocity limit that results -
[
in containment vessel damage equivalent to the damage it _would sustain during the impact test defined above. The application shall perform sufficient test and-analyses, specific to the test package characteristics, to support the selected impact velocity. The applicant shall also select an appropriate target design and perform l
supporting analyses verifying that it is effectively unyielding to the test package
[
impact.
I
- Containment: During and after the specified testing, the packaging shall not release p
- more than an A2_ quantity of plutonium per week.
This chapter presents the results of studies performed to assess the feasibility of'
~
conducting a PAT package Controlled Test by impact on an unyielding surface of an.
l-
- equivalent impact velocity. ;(A~ methodology'for estime:kg the equivalent velocity-L j<
q
...-.-.-.,_L-,
.,o-,-.,-_.,..
was explored during Phase One studies !Ref. 8].) A survivable PAT package model was developed by computer simulations of package model impacts in accordance with the test criteria outlined above. This was followed by physical impact tests with scaled package models and subsequent correlation and benchmarking of testing
)
results with the analytical method. Finally, an assessment is made of the achievability of deriving an equivalent package impact velocity.
2.1 Computational Analyses Package impact studies were conducted during the early '60s to evaluate the effectiveness of cushioning materials (Ref. 9]. The dynamic impact behavior of shipping containers on various types of impact targets was also studied for the nuclear industry [Refs.10-13). An impact velocity ratio or multiplier (the impact velocity on a yielding target to the impact velocity on an unyielding target that provides equivalent packaging damage or deceleration) was defined and tabulated in Ref.13. As pointed out in Ref. 8, these velocity multipliers become increasingly unreliable as the package impact velocity increases. At relatively high impact velocities, the package design characteristics become predominant in the behavior of the impact energy distribution and package deformation. Simple energy-balance estimates are no longer adequate to account for energy dissipation and package damage. A numerical modeling method capable of handling large plastic deformations, as well as proper representation of non-linear material behaviors, is necessary to adequately predict package dynamic behaviors and damage. The computational techniques that were applied for evaluating the dynamic behavior of package model impacts is described below.
2.1.1 Methodology In applying the computational techniques to a package subjected to high-velocity impacts, a methodology for identifying the survivability of packages as well as categorizing their damage level is required. The methodology reported in Ref. 8 has I.
been demonstrated to be effective for assessing package survivability and associated damage level. In that report, the maximum effective plastic strain produced in a package outer shell was used as an indicator to judge whether the package would l
survive the impact on real and unyielding surfaces. Based on experience, an effective local plastic strain upper limit of 60% was chosen for the package outer shell which is made of ductile stainless steel (e.g., type 304). At higher strain levels, failure of the shell may occur. As demonstrated in Ref 8, the plastic strain predicted by computer analysis correlates well with physical test results.
For a containment damage indicator, the maximum deceleration of the packcge i
L containment during impact was chosen. This quantity is meaningful because it j
relates directly to stress levels below the elastic limit in the package containment l !
~_
-~
.~
vessel. These quantities are adopted in the current analyses for evaluating packaging survivability.
Extensive computational analyses were conducted during Phase I studies [Ref. 8] to examine the survivability of a typical PAT package containing wood energy absorber. The conclusion reached in that study was that a package of typical I
construction would not survive an impact at the conditions specified in Ref 2.
Therefore, packages with a different construction and perhaps using an energy-absorbing material other than wood should be considered for high-impact-velocity applications.
In developing a package that would be able to survive high impact velocities, the following guidelines are recommended:
- 1. The stress intensity limits in the containment vessel should comply with Regulatory Guide 7.6 requirements [Ref.14] (i.e., maximum stress developed in the containment vessel must be below the elastic limit of the vessel material).
- 2. All critical structural materials (e.g., containment vessel) should comply with Regulatory Guide 7.11 requirements [Ref.15) (i.e., materials should have sufficient structural toughness to withstand localized cracking in the event of impact).
- 3. The maximum effective plastic strain produced in the outer shell of the package should be less than 60% (to ensure the integrity of the package).
A difficulty in modeling wood as an energy-absorbing material is that wood I
properties are anisotropic. Also, wood exhibits a buckling and splintering phenomena which is difficult to analytically model for impact conditions. Thus, only isotropic or nearly isotropic materials for use as energy-absorbing materials should be considered for high-velocity-impact packages. Candidate materials are I
grout (a special mix of cement and fine sand), high-density fcam, and resin. These materials have nearly isotropic properties which are preferred for obtaining reliable analytical predictions.
As explicitly stated in Ref. 2, an applicant may conduct a package Controlled Test on an unyielding surface at an impact velocity that is equivalent to 282 m/s impact velocity on the specified surface. Reference 8 presented a methodology to relate this
' equivalent' impact velocity impact on the specified yielding surface. Figure 2-1 is a graphical presentation of the ' equivalent' impact velocity concept. The same concept is adopted for the Controlled Test studies.
?
i 4
l
t 2.1.2 Computer Models Finite element analysis computer codes capable of simulating high-velocity-impact dynamics and large material and geometric nonlinear behaviors are required to accurately predict package impact events, The finite element codes selected to simulate dynamic events of package impact are DYNA 2D [Ref.16) and DYNA 3D
[Ref.17). DYNA is an explicit finite element code for analyzing the transient dynamic response of solids and structures. Available element formulations include one-dimensional truss and beam elements, two-dimensional quadrilateral and triangular shell elements, and three-dimensional continuum elements. Many material models are included in DYNA to represent a wide range of material behaviors. DYNA also has a sophisticated contact interface capability to handle arbitrary mechanical interactions between independent bodies or between two portions of one body. Over the last ten years, DYNA has been used extensively at LLNL and in industry. It has been applied to a wide spectrum of problems, many 3
involving large inelastic deformations and contact. The code has been benclunarked against many textbook problems (Ref.18) as well as with different dynamic applications.
After selecting a suitable finite element code for analyzing package impacts, the next task was to construct a generic representative PAT package according to guidelines given in Section 2.1.1. Following a number of initial iterations, a simple three-layer model was developed as shown in Fig. 2-2. The model contains an outer structural t
i shell, a layer of energy-absorbing material, and a structural container representing the containment vessel for transporting the radioactive contents. The model, though simplistic in nature, includes all the essential structural elements of a PAT package. The outer shell is made from ductile stainless steel. The inner containment is constructed as a nearly rigid high strength component. Thus, the kinetic energy from impact is expected to be absorbed mostly by the energy-absorbing material layer, and to a lesser extent, by permanent deformation of the stainless steel shell. The finite element representation of this three-layer representative package is also shown in Fig. 2-2. The model was generated using the two-dimensional mesh generator MAZE [Ref.19). The model takes advantage of the axisymmetric geometry of the package.
In the analysis of the representative package impacting on an unyielding surface, the unyielding foundation modeled in the DYNA code as an unpenetrable stonewall.
In the analysis of package impact on a yielding surface, the finite element mesh representing the yielding surface was generated using the mesh generator MAZE.
Dimensions of the yielding surface are chosen so that a condition of no stress wave interference due to wave reflectiors at the boundary throughout the duration of impact was ensured.
.s.
l-
l An important element in analytical modeling of package impact is a computer code's ability to accurately model material behavior. Materials are characterized in the DYNA code by choosing one of the available material models. Some material models require an equation-of-state to represent the pressure-volume-energy relationship in a material that undergoes large plastic deformation.
4 Two material models, which are available in the DYNA code, were used for modeling behavior of geological materials such as grout. The first model was used for preliminary analyses of the grout impact surface. The second model was used for analyses of the package model and final analyses of the grout surface. A brief description of each model is given in the following.
1.
Soil and crushable foam model:
This model is base on the formulation originally suggested in Ref. 20.
~
Volumetric and deviatoric responses of the modeled material are treated independently. The volumetric response is computed based on a user-defined volumetric strain-versus-pressure curve, which can be nonlinear. The deviatoric part of the response is calculated using the conventional plasticity theory with a radial return. The yield stress o, at any pressure P, is given by the y
expression G = [3 (a + a3 P + a2 P ))w 2
y o
where ao,ai, a2 = material constants that characterize the yield-versus-pressure relationship.
There is no strain hardening in the model, so the yield stress is completely determined by pressure. This model is useful as a simple represantation of pressure hardening, where the deviatoric yield strength cf a material increases as the hydrostatic pressure increases. At any constant pressure, the deviatoric behavior is elastic perfectly-plastic.
2.
Concrete / geological material modeh This model was developed to incorporate more features and thus offers more versatility in the modeling of geological and' concrete materials. It has the capabilities of modeling strain rate effect on the yield strength via the use of a load-curve multiplier. Material damage and failure phenomenon that is common in materials such as grout can also be modeled through the use of a two-curve concept. The two yield versus-pressure curves are defined as upper or undamaged curve and are represented by P
%=a+o 31+3P 2.
where o
= the material yi, I stress at the undamaged state, max P = pressure, and ao, ai, a2 = material constants that characterize the yield-versus-pressure relationship at undamaged state l
and the lower, or failed (damaged), curve, is represented by P
Utdled
- 80f + ag + a2P 4
j where crailed = the material yield stress at the damaged state, and aof,atf,a2 = material constants that characterize the yield-versus-i pressure relationship at damaged state.
By defining those two curves and defining an appropriate scale factor, y, versus the effective plastic strain, Ep, in this material model, as used in the following fashion
Ufailed + U (U
- Ufailed )
- Uyield max l
either a hardening or a softening phenomenon for the grout or concrete material can be described according to the amount of plastic strain levels produced in the material.
The pressure-volume relationship of the material is treated independent of its l
deviatoric behavior and is described using the equation-of-state type such as tabulation with compaction, or saturated, or air-filled, porosity formulation.
l l
A third model, titled extended two invariant geologic cap model, may be more appropriate for the grout but it was not used because there was insufficient data to define all the modeling parameters. This model is formulated in terms of the invariants of the stress tensor-the square root of the second invariant of the deviatoric stress tensor, J20, and the first invariant of the stress tensor, Ji. The cap model consists of three surfaces in J20 -J1 space: (1) a failure envelope surface that is fixed in J20 - J3 space, and therefore does not harden; (2) a cap surface that is 4
controlled by the hardening pcrameter x; and (3) the tension cutoff surface. One of the major advantages of the cap model over other classical pressure-dependent plasticity models is the ability to control the amount of dilatency produced under shear loading. Another advantage of the cap model over other models is the ability
_ to model plastic compaction.
Two other material models-an isotropic-elastic-plastic-hydrodynamic material model, which uses a tabulated yield stress versus effective plastic strain curve, and a i
j power law isotropic plasticity material model, which expresses yield stress as a
- function of plastic strain-are ideal material models to use in describing the,
1 nonlinear strain hardening behavior exhibited by materials such as stainless steel
~
outer shell or containment vessel material.
2.1.3 Simulated Impacts Computer simulated impacts were performed using the representative package.
Figures 2-3 and 2-4 show the axisymetric finite element meshes used in the simulation of package impacts on unyielding and yielding surfaces, respectively.
The study was limited to only one energy-absorbing material (grout) and only one type of yielding surface (soft rock). This type of yielding surface closely simulates the surface conditions specified in Ref. 2, The objectives of conducting this series of package impact simulations are: (1) to find a package configuration that would survive both impact scenarios, which are impacting on an unyielding surface at velocities above 129 m/s (422 ft/s) and impacting on a soft rock surface at 282 m/s; and (2) to optimize the configuration of this survivable package in terms of the thickness of outer shell, total weight, and the suitability of the containment vessel structure with respect to its maximum deceleration level developed during impact. Table 2.1 exemplifies one such final round of effort in the search for an optimized survivable package.
l In all package impact simulations, the grout material properties were approximated using the parameters given in Ref. 21, and its behavior was simulated using the concrete / geological material model described in Section 2.1.2. The yielding surface approximation was made using the soft rock properties reported in Ref. 8 and was simulated with the soil and crushable foam material model.
Figure 2-5 shows the predicted deformed shape of the package impacting on an unyielding surface at 137 m/s. Figure 2-6 plots the plastic strain contours of the outer shell of the package. No plastic strain was developed in the containment vessel. Figure 2-7 shows the deformed shape of the package impacting on the yielding surface at 282 m/s. Figure 2-8 plots the plastic strain contours of the outer shell of the package. Again, as with the case of package impact on an unyielding surface, no plastic strain was developed in the containment vessel and therefore no permanent deformation was produced.
2.1.4 Model Design for Benchmark Tests 4
Based on computer impact simulations and available material data, a PAT package model that could survive the required impact conditions was developed. To verify _
the adequacy of the computer simulations, physical benchmark tests are necessary.
Since it is imptr.ctical to benchmark test a full-scale model, a scaled-down model that captures di the important characteristics of impact, was tested. Thus, geometric scaling laws were applied to developing a scaled package test model. Appendix A,
er
~~~
e
.=
1 Table 2.1. Computer simulation results of the generic package impact event Package outer Package Impact vel.
Impact Contain vessel shell plastic config.(cm)
(m/s) time (ms) peak G strain (%)
Package end impact on unyleiding surface 63.5(dia)x112(lg) 137 2.8 11,172 62.7 4
x1.02(thk) 152 2.9 -
12,534 71.5 63.5(dia)x81(Ig)*
61 2.0 4,946 18.1 x1.02(thk) 91 2.1 8,359 32.3 137 2.2 11,840 62.6 63.5(dia)x81(Ig) 137 2.0
-12,723 51.8 x2.03(thk) i Package end impact on yielding surface (soft rock) 63.5(dia)x112(Ig) 282 8.0 9,765 52.8 x1.02(thk) 63.5(dia)x81(Ig) 282 7.0 17,635 47.8 x1.02(thk)
Package side impact on unyielding surface 63.5(dia)x81(lg) 137 1.8' 22,397 29.6 x1.02(thk)
' candidate reference package model discusses the essence of geometric scaling laws. Figure 2-9 shows the final test package design. All dimensions were scaled from the conceptual package according to the scaling laws. A flat aft end for the test package was used because simulations on the conceptual package showed very little deformation at the aft end from impact. Also, the flat aft end construction provides best geometry for propelling a test package. An aluminum ball was mounted inside the center container, I
simulating the package containment vessel, to obtain passive measurement of peak deceleration during impact.
4 i
Computer simulations of test modelimpacts were made before the physical tests were conducted. Figures 2-10 and 2-11 show predicted test model deformations pattern when impacting on an unyielding and on a soft rock surface, respectively.
2.2 Benchmark Tests A testing program was included with the Controlled Test studies to obtain computational benchmarking data. The objectives of the tests were: (1) to obtain data to improve the mechanical behavior models for grout, and (2) to demonstrate equivalence of impacts on unyielding and soft rock surfaces. Material property data were obtained from mechanical tests on grout samples and from deformations of package models subjected to impacts. Equivalence tests were conducted by impacting package models on unyielding and soft rock surfaces.
2.2.1 Grout Property Measurements Mechanical property tests were conducted on grout samples that were prepared according to the formulation given in Appendix B. Pressure-volume and triaxial j
compressive strength tests were conducted using standard test machines and following standard test procedures (e.g., ASTM D2166-85, D3148, and D2664-86). The samples were cast in 5.3-cm-diameter by 10.6-cm-long plastic sleeves and cured for specified durations before they were tested. Prior to the tests, the samples were removed from the sleeves and machined to approximately 5 cm diameter for matchup with test fixtures. Table 2.2 is a list of the tests that were conducted.
Results of the laboratory grout property measurements are given in Appendix C.
2.2.2 Impact Tests Package test models were impacted on unyielding and simulated soft rock surfaces.
The models were propelled by the 155-mm powder-driven gun shown in Fig. 2-12.
Test models and bags of black gun-powder were loaded into the gun breech. The gun is located at the Lawrence Livermore National Laboratory Site 300 Test Facility and is supported by personnel bunkers and control and instrumentation systems.
Test Model Figure 2-13 shows the test model construction. It is designed to be compatible with the 155-mm gun and to include the materials and basic construction of a PAT package. The principal components include (1) a center assembly, which represents a plutonium container, (2) an outer shell, and (3) grout impact-energy absorber. The center assembly shell is maraging 300 steel, which has a yield strength of.
approximately 1,380 MPa (200 ksi). This material was selected so that, during the impact tests, plastic deformation of the shell would not occur. The center assembly also includes a 1.22-cm (0.48-in.) diameter aluminum ball and holding screw for deceleration measurement. (The ball is described later in this section.)
Table 2.2.
Summary of pressure-volume and compressive strength tests conducted on grout samples Cure time Confining pressure (days)
(MPa)
(ksi) 14 0
0 14 138 20 14 207 30 m 21 0
0 28 0
0 28 35 5
28 138 20 28 207 30
- 3 samples tested 2 samples tested 1
4 The outer shell is fully annealed AISI type 304 stainless steel. A commercial pipe cap and pipe section were welded together to form this shell Prior to filling with grout, the shell was annealed and machined to a constant wall thickness. A back cover plate was welded to the cylindrical shell after the grout was added and cured.
The grout energy absorber was formulated as described in Appendix B and aged for
.1daysm when the model was tested. Two pours were required to fill the model with grout. During the first pour, the model was positioned vertically nose down and the center assembly was held in its correct position by a special fixture. Grout i
was added untilits level was slightly below the back end of the center shell. After the grout had cured for approximately two days, the fixture was removed, the l
aluminum ball and holding screw were installed, and the remaining volume was filled with a new mix of grout. When the age of the first grout mix reached approximately 25 days, the back cover plate was installed and welded to the outer i
shell. A small gap developed between the back plate and grout due to grout -
l_
shrinkage. This gap was filled by pouring an epoxy compound through holes in the back plate. This process ensured that no gaps or volds existed in the models when j
they were tested.
Impact Surfaces An impact surface having nearly unyielding properties for model impact tests was assembled from available materials. As shown in Fig. 2-14, a 5-cm (2-in.) thick steel plate backed by an 8.2 Mg (18,000 lb) block simulated an unyielding impact surface.
l This assembly provide a mass ratio of surface to test model of approximately 1,100:1.
U) Approximately 90% of maximum compressive strength is achieved at 28 days cure time.
! l l
l
- - -,, ~
- -,, ~ -, ~
4 An impact surface simulating the soft rock spccified by the Controlled Test criteria (Ref. 2] was provided by a 1.22-m (4-ft) cube of grout aged for 13 days. The grout block mass was approximately 3.6 Mg (8,000 lb) and was backed by the 8.2 Mg block as -
shown in Fig. 2-15. The ratio of unconfined compressive strengths of the grout and the specified soft rock is approximately 1.09:1. A corresponding ratio of bulk densities is approximately 0.88:1. A new grout block was used for each test.
Diagnostics Measurements of model impact conditions were impact velocity and model center-assembly deceleration. Other diagnostics included post-test photographs and radiographs of the models. In addition. high-speed motion photography of selected tests was made to view the impact events. Frame rates were approximately 8,000 and 2,000 frames /s.
Impact velocities were measured using contact pins placed near the gun muzzle.
Three pairs of pins spaced one meter apart were contacted by models while in flight.
Three timers connected to the pins provided three sets of time-distance data for calculating model velocities.
The aluminum ball in the model center assembly was used to measure peak deceleration of the assemblies during impact. After the impact tests, the diameter of a flat spot on each ball was measured. The spot was created by its interaction with the flat surface inside its container. (The spot diameter is related to the peak force the ball exerts on the container during impact deceleration.) Knowing the deceleration force (F) and the ball mass (m), a deceleration value (a) was calculated using Newton's equation F = ma. Fully annealed AA 1100 aluminum was selected as the most suitable material for the balls. They were a standard fabrication item obtained from a commercial supplier.
Before the model impact tests were conducted, laboratory calibration tests were conducted with the aluminum balls to determine a force versus spot-diameter relationship. Two test methods were used, dynamic and static. In the dynamic tests, the balls were shot from an air gun, over a range of velocities, onto a piezoelectric load cell. This test method developed mechanical strain rates in the balls that were approximately equal to rates that occurred during impact tests. In the static tests, the l
balls were compressed, over a range of forces,in a mechanical test machine. The results of the two tests are shown in Fig. 2-16.
Data from the dynamic tests exhibit some uncertainty as indicated by the scatter of data points in Fig. 2-16. The source of tbc scatter is electro-mechanical noise in the measurement system. As shown in the figure, data points for the static tests have much less scatter and exhibit a smoother trend. The general difference between data points from the dynamic and static tests can be attributed to differences in strain rates. The amount of difference is expected for the ball material and the experienced
- strain rates.
i l
! l F
Some prior studies on strain characteristics in impacted balls is reported in Ref. 22.
A force-spot diameter relationship is developed and expressed as F = 0.75 n o (spot dia)2 y
where oy is the yield strength of the ball material. A curve of this equation is included in Fig. 2-16 for oy = 58.6 MPa (8.5 ksi), which is a reasonable value for the aluminum balls. The value of o for annealed AA 1100 aluminum is approximately 34.5 MPa (5.0 ksi)y. However, work hardening of this material will occur when it is deformed as in the balls and cause oy o increase to approximately t
58.6 MPa. Thus, the curve in Fig. 2-16 is used to determine deceleration values from measured spot sizes.
l Test Plan l
Before conducting model impact tests, calibration tests with the 155-mm gun were conducted to determine powder load versus model impact velocity relationship.
The effect of model position in the gun barrel on velocity was also determined. A special model, which was identical to the impact test models except that it did not contain a center assembly, was constructed for the calibration tests. Several calibration shots were made at velocities ranging from approximately 120 m/s to 270 m/s. A calibration model was shot into soft sand and recovered for reuse in succeeding calibration tests.
4 Test model impact tests were conducted in two series. The first was onto the steel plate surface, and the second onto the grout blocks. Approximately one month between series was scheduled to allow time for review of test results from the first test saries before preceding to the next. Preparation of grout for the models and impact surfaces was scheduled so that the desired age was reached at the time of the tests.
2.2.3 Impact Test Results Six impact tests were conducted, four on the steel plate and two on the grout blocks.
The tests are summarized in Table 2.3. Peak deceleration values determined from recovered aluminum balls are also given. Using a turning lathe and parting tool, each tested model was dissected so that the holding screw could be removed easily from the center assembly and the bali 3 retrieved. When this operation was -
performed,it was noted tha'. the grout was compacted, fused, and unfractured as a result of energy absorptiori during the impets.
In the first three tests (~ 17,126, and 141 m/s), the impact angle of the models was approximately 5 to 20 degrees from perpendicular. When the high-speed motion film was viewed,it was noted that the model tilt occurred after exiting the gun barrel. There was an indication that non-uniform gas leakage occurred around the model while it was moving in the barrel, which could cause the tilt. This,
v
?
Table 2.3. Summary of model impact tests Impact velocity (2)
Deceleration Surface (m/s)
(ft/s)
(1000g)
Steel 117 383 58 Steel 126 414 63 Steel 141 464 81 i
Steel 157 516 65 4
Groutu) 206 676 54 Grout 288 945 78 U) Approximates the impact surface properties of the worst case aircraft accident crash site [Ref.1).
(2) NURECr0360 requires 129 m/s on an unyielding surface (Ref. 5],
Controlkd Test criteria requires 232 m/s on soft rock [Ref. 21 1
undesirable condition was corrected by installing a rubber O-ring in the plastic cap on the model. Succeeding impacts were perpendicular.
Photographs of the models after impacts at 157 m/s (516 ft/s) on the steel plate and 288 m/s (945 ft/s) on a grout block are shown in Figs. 2-17 and 2-18. Radiographs of l
these models before and after the impacts are shown in Figs. 2-19 and 2-20.
Photographs of the grout blocks after modelimpacts at 206 m/s and 288 m/s are shown in Figs. 2-21 and 2-22. As shown in Fig. 2-22, a large piece of the grout block broke off and fell to the ground. On viewing the high-speed motion film, it was noted that this occurred after the impact event. However, numerous fractures in the grout developed during the impact. Viewing of the high-speed film also revealed that the steel and grout surfaces did not noticeably move during impact.
2.3 Analysis of Test Results As summarized in Section 2.2, six package impact tests were performed, four on i
steel plate and two on grout target. Analysis of the test results yielded information for characarizing grout behavior under high-velocity impact.
2.3.1 Correlation with Analytical Model All the post-test models were carefully examined and no cracks were found in the outer steel shell (except for the model impacted at 126 m/s, where approximately l
0.5 cm of crack was observed at the weld joint between the package body and the cap). Also, no cracks were found in the containment vessel. In addition, no permanent deformation was observed in the containment vessel. Thus, a survivable' package model according to guidelines in Section 2.1.1 appears to be
j l
J Table 2.4. Adjusted vs. measured package deceleration for angled impact Angle of impact vel.
impact Measured Adjusted (m/s)
(degrees)
G G
116.7 5
58,000 57,779 126.2 17 63,000 60,247 141.4 23 81,000 74,501 157.3 0
65,000 65,000 feasible. However, the ' footprint' of the package impacting on unyielding versus yielding surface are very different, making it an uncertain measurement for establishing impact equivalence between the two surfaces.
The steel surface was not ideally unyielding due to deflection of the plate and associated movement of the backing block. Based on energy-conservation principles, an adjustment is made to the measured velocity to estimate an effective velocity for a non-deflecting surface. The most significant adjustment is for the 157 m/s impact, for which the adjusted impact velocity is 144 m/s.
Similarly, the measured deceleration for test packages impacting at an angle to the steel plate were adjusted using the measured impact angles listed in Table 2.4.
2.3.2 Development of Equivalent Impact Conditions The set of adjusted test package impact data was used te construct the ' equivalent' package impact conditions curve. The curve is shown in Fig. 2-23. Due to scarcity of the test data, straight lines were used to connect data points. If more test data points were available, both curves would probably curve upward. Following the procedure for equating two impact surfaces, an impact velocity of approximately 160 m/s on an unyielding surface is ' equivalent' to 282 m/s impact on the soft rock surface.
2.3.3 Correlation of Analytical Model Using Test Results l
A revised set of material constants were developed using data from laboratory l
measurements of grout properties (see Appendix C). This revised material 1 ;
representation was then used in DYNA code to correlate simulated impact results with that of test data.
Figure 2-24 shows the predicted test package deformation shape when impacting on the unyielding surface at 143.6 m/s. Figure 2-25 is an overlay of predicted package deformation shape impacting on unyielding surface at 143.6 m/s versus the post-test package deformation shape for impact on steel plate at 157.3 m/s, obtained from the post-test package radiograph. The comparison showed good agreement between the computer simulation and actual test result. The predicted maximum deceleration of the aluminum ball for this case is 58,851 G. The measured deceleration value from the aluminum ball was estimated at 65,000 G. The predicted maximum deceleration value is therefore about 9.5% lower, a difference within the experimental error.
The difference between analytical and test results is greater for the case of package impact on soft rock at 288 m/s. Figure 2-26 shows the predicted package deformation shape. The calculated results imply that the grout material may have softened during this impact. Therefore, the damage curve as described in Section 2.1.2 is invoked. However, no available laboratory measurements nor reported data relates the degradation of grout yield strength to plastic strain. A trial-and-error process is thus evolved in the determination of appropriate damage parameters Figure 2-27 shows the predicted package deformation shape using the latest set of (n,i )
p parameters (see Section 2.1.2 for definitions of 11 and E ). Eve. though the predicted p
i package deformation shape reasonably resembles that from the test radiograph, the predicted maximum deceleration of 141,356 G is 43% higher than the measured deceleration of 80,000 G. A lack of better grout property data restricts the capability of a computer model to make predictions of model impact behavior more reliable.
2.3.4 Equiuent Impact Condition Generated by Analysis Using the revised set of material parameters for the unyielding surface impact case together with latest set of grout damage (n, cp) parameters, equivalent impact condition curves are developed as shown in Fig. 2-28. The curves indicate that a test package impacting an unyielding surface at 190 m/s (650 ft/s) is approximately equivalent to impact on the grout surface at 282 m/s (925 ft/s).
2.4 Assessment of Controlled Test Studies By using a combination of computer impact simulation and model package impact testing, equivalent velocities for model package impacts on yielding and unyielding surfaces were estimated. While results can be improved with the availability of more laboratory grout property measurements, particularly in the area of grout damage, a methodology that can be applied to estimating velocity equivalence for use in the package Controlicd Tests has been demonstrated.
16 -
J
4 '
With the revised set of grout material parameters, computer simulations were conducted using generic package models shown in Fig. 2-2. Impact on unyielding and soft rock targets were studied. Figure 2-29 shows the resulting equivalent impact velocity curves for the generic package. The ' equivalent' impact velocities are found to be 185 m/s on an unyielding surface and 282 m/s on the soft rock.
2.5 Discussion In this chapter, a methodology for estimating equivalent impact velocities on yielding and unyielding surfaces is demonstrated. By conducting computer simulations and physical tests with a package model, equivalence in peak deceleration of the model container during impact was demonstrated. (Deceleration i
is related to stresses developed in the container.) Equivalent deformation of materials was not achieved because the impact dynamics on the two surfaces are different. Difficulties that would be encountered in determining equivalent impact velocities for an actual package can be avoided by conducting the Controlled Tests on simulated soft rock at the specified criteria (282 m/s on soft rock). Suitable soft rock impact surfaces can be constructed similar to those used for the benchmark tests (see Section 2.2.2).
An important part of conducting computer simulations of PAT package impacts is l
proper representation of material dynamic properties. This is especially true for energy absorbing materials included in packages. As demonstrated by the studies -
described in this report, the mechanical properties and failure characteristics of these materials are required, such as pressure-yield and pressure-volumetric strain relationships. These characteristics provide basic informatior_ of how energy-absorbing materials behave during high-velocity impacts. Each energy-absorbing material will have unique characteristics. Thus, a data set for each material is needed. To obtain the required data, laboratory tests on samples may be necessary.
By studying computer simulations of PAT package impacts, a model that survived the required impact conditions was designed. Also, by conducting physical impact tests with scale models, benchmarking of the computer material models was accomplished.
- i
~
,+
t c
.2
.E e
{
S-2 3,,
5 2
l f
5 j
a
.e f,a a
s A m e = ",
V y
0 1
impact velocity 4
Fig. 2-1 Graphical representation of equ vclent impact velocity concept i
i 3
/
l
\\ \\ \\ I 4 i
}
'"EY inner absorbin8 cont.
material l
llli
~~
' s Outer wrapper s
vi Fig. 2-2 Generic package model f 1
a h
_../
',e aa_
t 1:
ai
' sc M'
w Fig. 2 3 Finite element package model impact on un),:elding surf * 'e I
I s
D x
N
~
s
/
/
--)i Fig. 2 4 Finite element packase model impact on yleiding surface 19 -
1 i
~
1 i
j 1
\\1 4
I\\\\\\
i 1
I \\i s
i lll / /1 F
- M////
- r j
^
i i
sx
\\
u w
a n
-r
/
l' Fig. 2 5 Package deformed shape,137 m/s impact on unyleiding surface
{
i I
I-mint-lo 0.00C+00 mas (*le 6.2&C-01 i
i contour levels j
e* 1404C-Oi
]
be 2,09C 01 4
ca 3.13C-01 i
de 4.l?C 01
- 5.2FC 01 i
h t.
t 4
l
~
1 Fig.2 6 Package outer shell plastic strain contours,137 m/s impact on unyleiding surface 20-
., ;; y N
\\
N X
/
j
\\
N/
{
/
Fig. 2 7 Package deformed shape,282 m/s impact on sof t rock surface U"!.!'3:YSE:0?
contour levels ao 7.97E 02 be 1.000-01 c
2.39[-01 de 3.19C-01 3.99C 01
+
9 o'
Fig.2 8 Package outer shell plastic strain contours,282 m/s impact on soft rock surface i< r i
.: 90 jl:_
r(
l 4
l
(
j N
4 Y
i Li Fig.2 9 Finite element model of the test package
'I i
(
i
.4m l
il 1\\
14 t
i
\\\\
L
- '~
l Ili I \\1 i
Fig. 2-10 Deformed test package,137 m/s impact on unyielding surface r t Y
1 i
I e
i l
4 I
l I
iI i
mesurre
- m l
l qi [3. hNRMIrR11 gg j
h y-, t, 1
l.
[ r
.!f$ k~
. d /.
s.
R 1
l 1
1 t
Fig. 211 Deformed test package,282 m/s impact on soft wek surface s
v- -:4.,.n..?
.m.
~ ~,. n
',~ 4M'.f N: 3. ".<:.,f.$,,,. :.,. 1.'U Y m ~,.
v.
.. ~.
.~~~y u...
- 15 s.4(%ffi j ', i-
- '. (4.. e 1 ca.u.e. s. > w q C,
m,: +m H
.a -
u
. s e c.:.
p.
'x%..m.
- u u.-
.yi
-.~. s,.,.:..
. m, : :a....v.-Q'.w - ?.,_
~
y.
n%.."yQw'. ;.
,. 5. a'.:y:-
./~-
'~
7~ >r-4
-.,,' qe. ' : ' ~ y '.
.' %y
~.
.... a.
m 1 -:..t : f # $ % it. s *.%." ',t :
,:lT{. -
.,A e
.'I g.
X9,:
--(
j;;k o
- ., ~
h, YN
~~
~
h (a
,4-ma
..ig;s_@-er., s.
a c
,n)
.n _r f} ;;:.2:
W MA
- .:.9x ;
>f 17.,
N. n..
f' Im_.1$.k.
.d M A
~
Fig. 212 The 155 mm gun used to propel test models 23 '.
304 SST, fuhy annealed 707E'651 alum i
Lazan
- (1.350) l.
L(.109)
- p. w) :;i, :;i. :;i. *:;i.,.:li. ?.t./:;i..:;.i....h.* I
-9
- i.:
e p Grout i
5
,.:;i.:;k.j:;i.;::;. ' t :;'. ';';;;,.3.? t.* t.? /._*:.;,. ',,....
- ) :;i e 1100 alum 304 UST, L
.* t t.? t.* t.?
fuuy/
.'t
?I
' INU.*'e'M. *'t
.* t /* t.e'e.*'M'*54*N *0 Ii
- .,,* *t. * -
fuHy anneeled U.t t.* i annealed i
~.? tl.* 2 l.? il :
- l. :.*l?
.*r.
.:.r (
.:.ti,J.; sc..i.-
(6.181)
- S;* t ;
.*t.*i:.*e a,. -
(115)
. j *:.*l... l... Q
......t..*
t
- t..*
I
- I'.t. *
'p*.*p;t t p. t:.*p. p.*f. '.*p.*;;.. l *.*l.. Ij -
.9
.t
.r t M t.e t.e t.r t ;t t.s t. h i 1
p?.',;*.2,p?.E ;' 8.p?.'.;' 2.p. y* 2.p' #.;'*d..
- .' '. :.*l.
.*p.*
.?l.:.*
6
- 2 b Maraging 300 STL t '.
- l
........e.. 9;...;....... ;....p..;.. ; f
.?t.fR.tt att.*t.et.?t;fs. 94 a
r
[. -t. :..: :.s '. :. s:.s ' :. -: :.1 '.C.? ' :.e '.i.' '
4 o.ning
- (2.000) -
nuer I
(6.650)
Fig.2-13 The test model design f
i i
1 1
I
. '= p, * *..c.
4.
m
,.,~r~e
.s
+
~..
- u...
- y...
. ;;;.e., p y;.
,;. g.....c. w g.,
4 1
m e 4v,_
.:v. Q. '.
x
.g ? c..L.
i
., ~ ~
s, m
,04~;
s.
--duMAcAalf ad
- g f.
1
^ 9.~v-+ 3.R$
'%.').&
g*:ws -
)
.w s.
m
\\
+.
, $,, s, '.
.. A..
t x
R.t.
,s. 5.. w p;,,,.
x,.%n.
,1
.. c;..
4
.'... t}. t.r+ :, L,.. ~
7 ' %.
.;;!li'=,.
- % ~:
5-M
- w....
6.m. e..
.q e L.c.,v./..
p
.t.... -
o
> y %:-
.s,.
v.:p,M*.it*~%. ik
- n;'Afa...w w,' ~--
.w:.
o..
.-::y, -
...f.'.. ;gg....-
- yL
. l.w.... :
N
'..'e.,.
t *-
. c.;
.,Q n.
+.
- .~ '
~
w3.symg.
Fig. 214 The 5-em thick steel plate backed by an 8.2 Mg (18,000 lb) block l
l t l
o en.
g..
4,.
,7 ~ % ' : <'..,'.
~ ~,,.V..,.,.
1
. w.:. *.
3,.
.,.-+
yj.ce.a,-A.' *. ~..
h.q.:f,.g~h@,.,, =p ;r:.. -
g
%.s.A..*,.
u:.< w.
'.",.s.
r,.
s.
. s.,.. c ~.e.s. :~.,.,.
. s...a. 4._.
.., w
._,...,,y
,.V n..s x.
m
. =.. =
,5
..".?
1,'?
g
- . -it c.
g;._._._
.. =
=..
..g,. -t
...?,"..--=
/ I s
...,. ),%
'.r v..
.... +
. 3 f.-./
. :..~
~
Fig. 2-15 A 3.6 Mg (8,000 lb) grot.t block backed by an 8.2 Mg 08,000 lb) bloch
...,............. -......,...... - 200 c
1000 x
x s
900 x
,',',',',i',,o,,,,,,b 800 -
i i
x
,x 700 i, i,, o,,, ',x i
,, i i
.. 4. 4 q'=8;5 ksi', - M,'r,,'.-d',,5 90 o
,6 600
^
4-------- 100 m o
500
,,, i i i,
x,
... q. 4.. '.. ;. 4
- 1. ;..,
.p;. p t 4....... _
go S S 400
...:.. '.. L. ;. 4. '.g 'x. N.l. ' ' 4....... -
70 7 r
... j.. j..,.. t. j. j,..;. ' j.';. 'y } '...... - -
60 S l
8 y
- l
- xl l E
c5 300 a, w,,,',',,,,
.oO.9 u.
c3
,x, o
... s,.
4,... i, 2,. 1,. u, _,. u,.,. u, a,....... _ 40 o4 200 x,
......,.-s,.ss.ms.,........-
30 i i i i, x,, 6:atic force
.x,
l ', l l l l P : Dynamic force 20 t
1 100
~^
0.06 0.1 0.2 0.3 Spot diameter (in)
Fig 216 Calibration data for al..minurn balls used for accele:ation measurements 25 -
i J
1 j
- ~; - "M s
I i
- 3....
j l
l b.:,w - r
~ ~~
- / '
DMy y.
.g.,
- vt*
I
%?W m; ;.-
A M
N,., h,
%l.
',., R j
- ~
N 1 J.
4 1
>.6 7.Ner c
,,.f..; - - -
i l
1 Fig. 2-17 Test n adel after impact on steel pli.te at 157 m/s (516 ft/s) j
~~
(
(
(
)
l
[
l 1
n is m g
1 F
J til-i l
g.,,1f
__I
}_ 3 1
diI
~
ii e sa %.
- 4 y;
- , r, 8:-
p
- N,O:ifr(9,g,Au iii3 L.
%.~
1-I; "q' "'
.c m,.v V
A w
Fig. 2-18 Test model after impact on grout block at 288 m/s (945 ft/s) 26-
_ _ _, _ _ _. ~ _ _ _ _ _ _ _ _ _ _ _ _
e f Y;Yhj 'f(h f, "Q)sy;$;'*'Nihb;fDfi ' ~
Qf pir k' vr.+r *~rw s.CSd -mm " ' <
4, r
$$$isy d'kk".d ?#
- w Nh$f h $ hykhj
- fl#kt$5f} IND$hhfhp.(Y
. ';f$hf kr T
$is,($g($$eji I fh khh8 b k)$)AA$
m 4n =
eH 7sausmar.uw?gL.. uy&;a;,.s,m G
,y m 4,L naw e tc*wuf.w 7,
Fig. 219 Radiograph of test model after impact on steel plate at 157 m/s (516 ft/s) y B wlf ' 9$&L ' "4.;11
- ~ R. ?Ni(S!i?j 5 &O l? I gjf$e~fRA% dgq c *3ny f,pd}te{h ,c h i M ,M ##"*WW M M %%Idi IN ~ q! M $$3 NIfp.,k.,. I!;. c' jej W4% r r.,- q gues g m.tw(x ? ' u px y r AM M q '%.: k?s ";;k@ ". "d. Q ? m r h !.M 9 5fd&Bb d .d Fig. 2-10 Radiograph of test model,fler impact on grout block a ?SS m/s (945 ft/s) 27 -
9 ) [- k ] l J i dw f .g,, '? s 1 A i.! i .,x.,. ,, i .A.'[. ' M.-: ('.,'h E f ) \\ y n' y#,Q l 1,q 3 t lfr;Af }.
- .,.. r. ;,.
z. y . t. x t .. f,. % ', g* w
- .,,.n
,1 n ) )wln.Lg.'; w& . m.1 is ~ Fig. 2-21 Grout block after impact by test model at 206 m/s (676 ft/s) i
- 9.,; 7 : m.:.;,,,,..e '- y.
c., r e
- s..
w .ff; .'ty*' 7 g i ..r ,::. f,.. -..y.. i y ~ m.: < - w...W. w m,' s,c-~,b,5 : a
- -,~;
.a ~-s +.- , ', %c,t. 'A' 3R b
- ?-
' >... e. ,g. .yg~, Ly N% .,? Wi .$5.'. p,'s j' y7. '.. - ..z..:.,...., . v.,, , = - - a.:.. .+ . z. ~ 4 ;: p i :;._ ...; - :-c > .w l )s:W-,p -" ,e e - ei.; a - -- ,s ., ~; ec:. - ..m . F ' Tr= b. t c., 7_ - c a. ff.. -
- T i -s J,
y ./ -a MA ,,f: :9 jqg.. ;.g.g. f.. ?.'jfb i QN' ~ e.
- -c.
lifl.f::X Q,6 4,:." ~'- % 9. : -- W" f.%;}-lc*Y.L*:-w .rg:, 4,
- m... _.. ::s _
,. z-. l Fig. 2 22 Grout block after impact by test model at 2SS m/s (945 ft/s) l i. -
I i i 1 4 I i 80000 u 4 e l s 60000 l l Tarcet u e 1 l l L i l a steel 4oooo. 1 e s
- =
l e l 4 grout E i
- 5 20000 -
l E l e 3 e e U O + 0 100 200-300 t 4 Package impact velocity (m/s) { Fig. 2 23-Equivalent velocity curves for test package generated based on test data } 4 i I (, < / i a a i n n 1 4- . in tF itu \\, \\,1, \\\\R i i1 1 1 ) \\1 4 Y Fig. 2 24 Deformed test package,143.6 m/s impact on unyleiding surface - i 29 [
';JMgfgJ3?Mi4W WDMSLC4'i .?..;.g. ggr rm m.wsss+mntgdq .a.a t w y
- :tp; wa ww
< mr p . vn c ~
- a jqn;t-b.
wem
- i w r+4%ft s" 4
ity '. H q 1 r,
- y
.3 ( lyh h.- 9h---s 41 - e4 y- ,1, ce@w -p Engn.v t 3g m w. m w a-711. l ~ p MU
- kh' hj
!! { }
- h y {'
y S M:2~3EETFfb 'i E -c ~ T,nj. v.an + % m 9 f w DW5w r as pg =---y ,,h., &, $.K.. r. I '.,i h p" h.,5Y.Ulit,! .e l
- t I A hN
- N! %.-.ad((, hlb M&ttME Er--.. mygM(fgh Fig. 2 25 Overlay of computer simulation on radiograph results of test packar,e, impact on unylelding surface at 143.6 m/s EM 4
.i, i liite 1 i i! i i F* t'd '\\\\'.\\\\3 i 1< 1 a i 1 A, 1, ; 11 -i 8 11,.
- :M
! '. \\ \\ ' 1 b 4: lll; \\4
- H,',',
1 , i I ii ,/ sj ns-y ' / Y y ilhlYlYI -t y A 3 A / / p /\\ J - x,-{\\ -J.
J, p-
) ,- q /"g /~hh l Fig.2 26 Simulated test package deformed shape with no grout damage curve invoked,28S m/s impact on sof t rock surface .v.
= 'g i i n. i s g \\ q vw t iT i ~:[. [- I r 1 s __f %QQY ?T Fig. 2 27 Simulated test package deformed shape with grout damage curve invoked,288 m/s impact on soft rock 200000 u ] Tnrcet a 2c. i E-100000 - l l 0 steel u E l l 4 grout ,5 l 3 e g l l C I 0 0 100 200 300 Package impact velocity (m/s) Fig.2 28 Equivalent impact velocity curves for test package based on computer simulation. - - _ _. - -
20000 Tarnet i 1 l x 0 1 l m-unyleiding a. -i i 1MM-i + soft rock eu i E .5 l l 3 I i =o o- ^ i. O i i o 3co 200 300 Package impact velocity (m/s) Fig. 2 29 Reference package impact velocity curves generated by computer simulation 32 '-
- 3. CONTROLLED TEST METHODS The feasibility of conducting Controlled Tests that would comply with the criteria S ven in Ref. 2 is evaluated. Two candidate test methods are examined: (1) rocket i
sled and (2) aircraft drop. One available tcst facility and support organization for each method are reviewed for suitability. A primary requirement is that the test method must be capable of impacting a test package onto a surface of specified minimum hardness at a velocity of at least 282 m/s (925 ft/s), or onto an unyleiding surface at an equivalent minimum velocity. In addition, the package impact orientation (e.g., end, side, or oblique) must be controlled. Typical package weights are between 200 kg and 2,500 kg. The results of the evaluations are described below. 3.1 Rocket Sled Two types of rocket-sled test facilities located in the United States, cable pull down and track, are reviewed. The facilities are the property of a U.S. government agency and operated by a contractor organization. Any U.S. government agency or any organization sponsored by a U.S. government agency (including foreign organizations), may request the operator contractor to conduct rocket-sled tests. In the reviewed cable pull down facility, a rocket sled is used to pull test items onto a ground target from an aerial suspension cable. Impact velocity and orientation can be controlled. Maximum impact velocities are typically 275 m/s (900 ft/s) for a 360 kg (800 lb) test item and 245 m/s (800 ft/s) for an 1,100 kg item. The existing impact surface is an in ground,900 Mg, steel reinforced concrete block with a 10 to 20 cm steel surface plate that simulates an unyielding surface. Other impact surfaces must be fabricated. Sensors in a test item can be direct wired to recording equipment to make dynamic measurements during impacts. Track type rocket sled facilities having track lengths up to 1.5 km are available. Test items weighing 2.3 Mg (5,000 lb) can be accelerated to 600 m/s (2,000 ft/s). Required impact targets are not available, but they can be fabricated and positioned at the end of the tracks. A water braking system is available to stop the sleds before they reach the targets. Thus, test items are released from the sled before impact. However, impact velocity and orientation can be controlled. Telemetering systems are used to make dynamic measurements of a test item's response during impact. The review of this rocket-sled system indicates that it can be used to successfully conduct Controlled Tests. 3.2 Aircraft Drop The free-fall velocity of a test package dropped from an aircraft can become greater than 282 m/s (925 ft/s), which is required to satisfy the Controlled Tests criteria (Ref. 2]. However, this can be accomplished only if certain conditions are met. The -
release altitude must be sufficiently high and the ballistic numberm (BN) of the test mackage must be sufficiently large. Based on studies of the Cruise Altitude Drop Test j 1 ,Ref. 6], a UN greater than 10 Mg/m and a release altitude higher than 7 km 2 l (23,000 ft) are needed. These values are indicated in Fig. 31, which shows the 1 relationship between impact velocity and DN. The BN can be increased to a required value by increasing the test package weight and by decreasing the drag coefficient. A method that can be applied to reducing the drag coefficient is to enclose a test package in an acroshell as illustrated in Fig. 3-2. l The acroshell provides a streamline profile to airflow over the package and j maintains it at a desired orientation. For the design shown in Fig. 3-2, drag j coefficients less than 0.1 can be achievedm [Ref. 23]. To increase a test package weight, parasitic weights can be attached. For example, the weights can be of thin lead sheet aad they can be designed to have little effect on package response to impact. An aeroshell can also be made to have little influence on package impact response by constructing it of thin material, such as aluminum. l These materials have relatively low mechanical strength compared to typical package structural materials and would absorb relatively little impact energy. The feasibility of conducting a package impact test by an aircraft drop method has 4 been previously reviewed [Ref. 6]. A conceptual method for releasing a test package from an aircraft and resulting free fall trajectories were also developed [Ref. 6]. Furthermore, the study revealed that all needed services, equipment and technology are available to conduct such tests. A test range, operated by a U.S. government agency, that is suitable for conducting high altitude air-drop tests, and experienced in conducting such tests, was found. The range operator can provide all j instrumentation, safety, security, and other services needed for a successful test. In addition, review of a candidate test area in the range indicates that a soll hardness that would satisfy the controlled tests criteria [Ref. 2] is also available. Any U.S. l government agency, or any organization sponsored by a U.S. government agency j (including foreign organizations), may request air drop tests at the range. A U.S. government agency that can provide suitable air-drop services was also found. The agency has extensive experience with designing and constructing air-drop systems, such as drop platforms, release systems, instrumentation, and i parachutes. In addition, the agency can provide a suitable drop aircraft, such as a C 130 aircraft, and a properly trained flight crew experienced in air drop flights. Like the other U.S. agencies, any U.S. government agency, or any organization sponsored by a U.S. government agency (including foreign organizations), may request air drop services. p
- Cd, where W= object weight, A = object profile area facing direction of O Ballistic number = W/A p
motion.and Cd = drag coefficient. W Drag coefficient for air-dropped test packages cannot be accurately determined from available data. 4 Values for a specific design should be measured before conducting a drop test. 34
A i i i A conclusion of the study is that Controlled Tests can feasibly be conducted by the l air drep method. 400 Drop aWtude 7 km 1 j imDact elevation Controlled test Okm j \\ s 3 ,.m 300 e Z 2 sm 3km y c S 200 i N 15 -j i E e - 100 l 0 t 0 5 -10 -15. i 1-i Ballistic number (Mg/m2) 4 l Fig. 31 Impact velocity of air-dropped objects released at 7 km (23,000 ft) with respect to ballistic number and impact elevation 4 l Aero shell heightTest package /-Fin (4) W / XT
- 1
-d l L-l- 2d l. End drop configuration Fin (2) Aero shell 2d-Test package s /,,,,,J. Weight (2) N }d + l 4 ..~~~ Side drop configuration i Fig. 3 2. Conceptual acroshell design for PAT packages to be impact tested by air- - drop method. 35-z.- -....
i
- 4. CRUISE-ALTITUDE PACKAGE DROP TEST 4
Section 5062 of Public law 100 203 [Ref. 3) requires a test PAT package to be dropped from the maximum cruising altitude of the cargo aircraft. A proposed Controlled Test method, described in Section 3.2, also involves dropping test packages from I aircraft. Thus,if these two tests were combined into one test, the number of tests would be decreased. 4.1 Cruise-Alt. ude Drop and Controlled Test Impact Conditions The maximum cruising altitude of cargo aircraft that would typically be used to transport PAT packages from a foreign nation to a foreign nation through U.S. airspace is approximately 13 km (43,000 ft) [Ref.1). Objects released for free fall from j such an altitude would typically achieve their maximum velocity before impacting i the ground. This is indicated in Fig. 4-1 which shows free fall velocity relationship to altitude and ballistic number (UN) for objects released at 13 kmH). The curves show that vele "" decreases after it reaches its maximum value because air density increases with .c asing altitude. Typical DN values for packages without special drag reducing 4.ichments is between 2 and 4 Mg/m2. As indicated in Fig. 4-1, packages having BN values in this range would have an impact velocity less than required for a Controlled Test. Thus, a typical PAT package subjected to impact by the air-drop method that satisfies the Controlled Test criteria (Section 3.2) would experience conditions more severe than it would in a Cruise Altitude Drop Test. Therefore, a package air drop test satisfying the Controlled Test criteria can also satisfy the Cruise Altitude Package Drop Test criteria, and the two tests can feasibly be combined into one test. Finally,it appears that the required Cruise Altitude Package Drop Test can be conducted in such a way that it could also be demonstrated to an independent review panel that stresses in a package container would be greater than those that would occur in an actual Aircraft Crash Test. 1 4.2 Test Range Evaluation Test range and support services described in Section 3.2 for Controlled Tests i conducted by the air-drop method are also suitable and availabic for conducting a Cruise Altitude Package Drop Test. H) Analyses of air-dropped o}ects were performed for the development of criteria for Cruise-Altitude Package Drop Test (Rd. 51 36
4 400 Drop altitude - 13 km l 350 m W 300 ~ t 1 & 250 1 8 ) 200 k,g,, BN-4 ~ ' BN - 6 150 'BN-8 100 0 5 10 15 Altitude (km) Fig. 41 Free fall velocity of objects released at 13 km (43,000 ft) - 37
- 5. AIRCRAIT CRASH TEST An actual crash test of a cargo aircraft fully loaded with full scale samples of containers with surrogate material is one of the requirements stated in the Public j
Law 100 203 [Ref. 3] for the certification testing of a proposed PAT package. The crash test must replicate or exceed the conditions surrounding the crash of PSA Flight 1771, defined by the NRC to be the worst case accident [Ref. 24). A comprehensive study of the Aircraft Crash Test was conducted during Phase One of this Project (Ref. 6). The study considered three issues: (1) key Aircraft Crash Test parameters needed to replicate the crash of PSA Flight 1771, (2) identification of problems that might threaten the integrity of test aircraft prior to impact, and (3) considered the feasibility of converting the crash test aircraft for remote operation to achieve the specified impact conditions. The following sections reiterate some of the important conclusions reached in the study regarding each of the three key issues. 5.1 Alternative Test Aircraft A brief study was conducted in Phase One to determine if there are significant differences in the environments that a PAT package would be exposed to in a crash of a Boeing 747 and a Boeing 707 aircraft. The study centered around three main aspects:(1) loading arrangement of the packages;(M unique aircraft components that could cause puncture or large deformations of the pakage; and (3) hardening of the impact surface resulting from aircraft impact. From analyses of probable physical arrangements of packages,it was concluded that there is little lateral packtige interaction [Ref. 25). As a result, a single linear series of packages on the aircraft is subjected to essentially the same loading as multiple linear arrays. Analyses further showed that the front package is subjected to the highest load at impact, regardless of the number of packages in line behind it. Loading for the front package is 46% higher than for the second package in the case of hard rock. impact and 25% higher for weathered rock impact. Therefore, from the aspect of loading arrangement on the aircraft, the crash environment equivalence will exist between any likely cargo aircraft and any test aircraft that can accommodate at least one package. A careful examination of the structures of Boeing 747 and Boeing 707 aircraft finds that there are substantially more massive components present in the Boeing 747.' However, when a cargo aircraft is specified,it should be possible to assess its massive components of concern and install taose in a physically equivalent (but not operational) manner on another type of aircraft selected for the Aircraft Crash Test. Equivalence of crcsh environment should be achievable. A firm conclusion can be made after the cargo aircraft is specified and compared with a candidate test aircraft. 38. -
1 The issue of possible impact surface hardening caused by the fuselage on impact was examined in Ref. 25 using simplified models of Ilocing 747 and Boeing 707 aircraft. The results indicated that there is not a substantial difference in hardening caused by impact of either aircraft. While the Boeing 707 impact showed slightly more soil compaction, slightly higher soll pressure, greater penetration, and roughly equivalent deceleration, no measurable difference in package impact response is expected. A conclusion reached from the crash environment equivalence study is that it should be practical to achieve crash environment equivalence in a test aircraft that is not the same ty, as the spelfied cargo aircraft. 5.2 Flutter Avoidance As reported in Ref. 26, the allowable flight envelope of the PSA Flight 1771 aircraft was exceeded by a substantial margin. This raises a crucialissue of whether the test cargo aircraft could remain intact until the required impact is achieved or could it suffer an uncontrollable Hutter condition that would damage the test aircraft structure prior to impact. It is important to ensure that the test aircraft has a high probability of impacting at the required velocity without any structural damage. A study to determine the Hutter-free regime boundaries for the test aircraft will be necessary to obtain this assurance. Presented in Ref. 27 is a status report on aircraft Outter and related subjects as well as some current industrial practices used to reduce aircraft Outter. The report includes: (1) a general description of the aircraft Outter phenomenon, its structural representation, and some of the commonly used unsteady oscillatory aerodynamic loading approximations; (2) a brief survey of the flutter equation solution techniques; (3) some of the main aircraft flutter characteristics and prevention considerations; and (4) an outline for future work on the aircraft flutter phex:nenon. An important conclusion given in this document is that detailed structural imbnation for an aircraft, such as dimensions and geometry, weight and stiffness distributions, control surface locations, external engines, etc., are required to analyze the aircraft's flutter characteristics. However, some or all of these aircraft data are proprietary to the manufacturer, which presents a possible obstacle to investigating aircraft Outter. This implies that calibration with appropriate aircraft manufacturers or modifiers must be established to secure relevant data necessary to conduct a successful flutter analysis and thus to ensure that the test aircraft remains intact until impact. 5.3 Flight Control Review The feasibility of modifying a crash test aircraft for remote operation to achieve the specified impact conditions was reviewed during Phase One. Three issues were i i }- I t i l examined: (1) the dynamic response of an aircraft and the ability to control the l aircraft in Olght so that the required conditions can be achieved; (2) whether the l control architecture provides for autonomous operation or remotely piloted (man-in the loop) operation; and (3) whether there is related experience or will new j technology be required. On the issue of dynamic aircraft response,it was concluded j that no new technology is rege'. red to Oy a large aircraft with a terminal homing guidance system. However,it may be necessary to augment the existing control systems (rudder, stabilizer, elevator, and alleron actuators) in the test aircraft in order to obtain sufficient control authority. On the issue of which. control architecture to select,it _was recommended that the final impact be entirely under automatic control on board the aircraft. A microwave beacon placed at the desired l impact point for a terminal homing guidance system in the aircraft appears to be a l reasonable approach. Lastly, there existed considerable experience with converting i numerous aircraft into remotely piloted vehicles (RPVs) over the last 40 years in l pilotless alght of large aircraft. Therefore, although there is only one opportunity to j perform the crash test, developmental and practice alght tests that will achieve all-l the specified impact conditions can be performed without crashing the aircraft, and j the goal of conducting an Aircraft Crash Test with an actual cargo aircraft should be achievable. I e + i i L i e 4 i i k- ; we rw-, s r m w ww-e rwe e 2 mm.- a v m+ c---*.swre---i. , e m - w e, w tw--,w.#-w,U,., y+,,E,--w e ww.m U.. w rs,A _.. +,-+-,ves-*m,3-w ew -,,
0 4
- 6. REFERENCES 1.
C.E. Walter, J.H. VanSant, and C.K. Chou, Draft Criteria for Package Drop and Aircraft Crash Tests - An interirn Report, UClD 21689, Lawrence Livermore National Laboratory, Livermore, CA (June 1989). 2. L.E. Fischer, J.H. VanSant, and C.K. Chou, Draft Criteria for Controlled Tests for Air Transport Packages, UCRL-ID-103684, Lawrence Livermore National Laboratory, Livermore, CA (August 1990). 3. United States Public Law 100-203, Title V - Energy and Environtnent Prograrns, Subtitle A - Nuclear Waste Arnendrnents, Part F - Miscellaneous, Section 5062 (December 22,1987). 4. United States Public Law 94-79, Nuclear Regulatory Corntnission Authorization Act for Fiscal Year 1976 (89 Stat. 413; 42 U.S.C. 5841 note) (August 9,1975). 5. United States Nuclear Regulatory Commission, Qualification Criteria to Certify a Package for Air Transport of Plutoniurn, NUREG-0360, Washington, DC (January 1978). 6. C.E. Walter, editor, Developrnent of Draft Criteria for Package Drop and Aircraft Crash Tests and Feasibility Review, UCRL-ID-103497, Lawrence Livermore National Laboratory, Livermore, CA (July 1990). 7. J.H. VanSant, L.E. Fischer, R.W. Mensing, J.C. Chen, T.F. Chen, C.K. Chou, J. Hovingh, and M.C. Wltte, Developrnent of Criteria for Controlled Tests for Air Transport Packages, UCRL-ID-104484, Lawrence Livermore National Laboratory, Livermore, CA (September 28,1990). 8. T.F. Chen, The Mechanical Response to Irnpact of a Representative PAT Package, UCRL-ID 104864, Lawrence Livermore National Laboratory, Livermore, CA (September 27,1990). 9. W.G. Soper, and R.C. Dove, " Similitude in Package Cushioning," ASME Journal of Applied Mcchanics, June 1962, pp. 263-266.
- 10. J.T. Schamaun and W.A. Von Rieseng nn, Effects of Target Rigidity on linpact Behavior of Shipping Containers, SAND 76-0207, Sandia National Laboratory, Albuquerque, NM (June 1976).
- 11. L.L. Bonzon, Final Report on Special Irnpact Tests of Plutoniurn Shipping Containers Description of Test Results, Sandia National Laboratory, Albuquerque, NM, SAND 76-0437 (February 1977).
41 -
= i \\ 4 Is
- 12. L.L. Donzon and J.T. Schamaun, Container Damage Correlation with Impact j
Velocity and Target Hardness, Sandia National Laboratories, Albuquerque, NM, ERADA contract AT(29-1) 789 (1976).
- 13. J.D. McClure, H.R. Yoshimura, R.B Pope, and R.M. Jefferson, Relative Response of Type B Packagings to Regulatory and Other impact Test Environments, j
Sandia National Laboratories, Albuquerque, NM, DOE contract DE-AC04-1 76DP00789 (1976). 1
- 14. United States Nuclear Regulatory Commission, Regulatory Guide 7.6: Design Criteria for the Structural Analysis of Shipping Cask Containment Vessels, U.S.
Nuclear Regulatory Commission, Washington, DC (June 1991). l
- 15. United States Nuclear Regulatory Commission, Regulatory Guide 7.II: Fracture i
Toughness Criteria of Base Material for Ferritic Steel Shipping Cask Contain:nent Vessels with a Maximum Wall Thickness of 4 inches (0.1 m), U.S. i Nuclear Regulatory Commission, Washington, DC (June 1991).
- 16. J.O. Hallquist, User's Manual for DYNA 2D-An Explicit Twc-ljimensional Hydrodynamic Finite Element Code With interactive Re:oning and Graphical Display, UCID-18756, Rev. 3, Lawrence Livermore National Laboratory, Livermore, CA (March 1988).
- 17. R.G. Whitley, DYNA 3D-A Nonlinear, Explicit Three Dimensional Finite Element Code For Solid and Structural Mechanics - User Manual, UCRL-MA-107254, Lawrence Livermore National Laboratory, Livermore, CA (May 1991).
- 18. S.C. Lovejoy, and R.G. Whirley, DYNA 3D Example Problem Manual, UCRL-MA-105259, Lawrence Livermore National Laboratory, Livermore, CA (October l
1990).
- 19. J.O. Hallquist, MAZE -An input Generator of DYNA 2D and NIKE2D, UCID-19029, Rev. 2, Lawrence Livermore National Laboratory, Livermore, CA (June i
i 1983). l
- 20. S.W. Key, HONDO-A Finite Element Computer Program for the Large Deformation Dynarnic Response of Axisymmetric Solids, Sandia National Laboratories, Albuquerque, NM,-Rept. 74-0039 (1974).
- 21. L.E. Fischer, C.K. Chou, T. Lo, and M.W. Schwartz, Structural Evaluation of the Shippingport Reactor Pressure Vessel and Neutron Shield Tank Package for l
Impact and Puncture Loads, UCID 21438, Lawrence Livermore National l Laboratory, Livermore, CA (June 1988). P L ..u. 4 l
- 22. C.ll. Mok and J. Duffy,'The Dynamic Stress Strain Relation of Metals as Determined from Impact Tests with a Hard Ball," Int. J. of Mech. Sci., Vol. 7,
- p. 355 (1965).
- 23. S.F. Hoerner, Fluid Dynarnic Drag (published by author,1965).
- 24. Nuclear Regulatory Commission, Selection of a Worst Actual Aircraft Accident for Murkowski Arnendtnent Irnplernentation (Public Law 100 203), NRC Memo SECY-88-344 (December 15,1988).
- 25. M.C. Witte, Structural Itnpact Analyses, UCRL ID-104576, Lawrence Livermore National Laboratory, Livermore, CA (April 10,1990).
- 26. C.E. Walter, investigation of the Ernpact Conditions and Crash Environinent of the PSA Flight 1771 Aircraft Crash on Decernber 7,1987, UCRL-LR-103735, Lawrence Livermore National Laboratory, Livermore, CA (February 28,1990).
- 27. T.F. Chen, Flutter Consideration of a Crash test Aircraft (A Review Article on Aircraft Flutter), UCRL ID-107239, Lawrence Livermore National Laboratory, Livermore, CA (February 28,1991),
m f 43-
l i 4 l ) APPENDIX A SCALING LAWS Presented in Section 2.4.1 is the construction of equivalent impact velocity curves using a reference" PAT package containing grout as the energy absorbing material. j The reference package dimensions are 63.5 cm diameter by 81 cm long and the steel shell thickness is 1.02 cm. The containment vessel dimensions are 7.62 cm diameter by 29.5 cm long Results from analytical studies of the reference package may be able to extend to reference packages of other dimensions by application of the geometric scaling laws. The same kind of extension can be made from scale test models to full-size packages. These prindples were applied in the Controlled Test studies described in Sections 2.1 and 2.4. i i 1 { A.1 Background i The use of scaling laws in structural dynamics studies has been explored by many researchers (Refs. A-1 to A-7 in Section A.3, References). Application of the scaling i l laws to scaling of geometrically similar structures subjected to large dynamic loads that cause inelastic material response and the restrictions on their use are found by cxatnining the scaling laws derived from dimensional analysis. Duffey (Ref. A-8) 4 has employed the Buckingham Tl theorem to generate 19 dimensionless parameters that must be equal for geometrically similar models and prototypes to have the same scaled response. In general, the relevant variables used mostly in structural dynamics are those that describe the geometry of the structure (e.g., length), the boundary conditions (e.g., impact velocity), and the response (e.g., time of i deceleration). Table A.1 lists all the pertinent variables. S is the scale factor. l Because each of the sixteen variables in Table A-1 can be ex' pressed in terms of three fundamental units (length, L; force, F; and time, t), they can be formed into thirteen independent dimensionless groups that govern the dynamic response of inelastic structures. This set of functional relationships describes the behavior of both the scaled model and the reference model (prototype). The variables in both systems must satisfy the i same functional relationships regardless of the system of units used. An examination of the dimensionless groups reveals how the variables must scale. For example,if the same materials are used, the stress and strain in the model and the prototype will be the same. This means that for the same impact velocity, the corresponding events for the scaled model occur a factor of S times faster (or slower) than in the prototype, and that the accelerations are also a factor of S greater (or less) in the model than in the prototype. This also means that we would like to see the gravity ( g ) increase (decrease) and the strain rate ( c ), the fractare toughness ( K ), e and the viscosity ( p ) of materials decrease (increase) by the same factor. However, these four variables do not scale for most of the known materials. In principle, the effects of these nonsealing variables could cause the model response to differ from s that of the prototype;in reality, these effects may not be as important as the scaling laws dictate. Table A.I. Scaling of Variables Variables Prototype Scaled Model Length L L/S' Displacement X X/S Strain c c Stress o o Time t t/S Velocity V V Acceleration a aS Strain rate c c/S Gravity g gS Solids Density p -p Young's modulus E E Rate constant t t/S Stress intensity factor K, Ke /S Fluids Density P P Viscosity p p/S Bulk modulus K K
- S is the scale factor A.2 Discussion Geometric scaling dictates that the gravitational acceleration in the model be S times the gravitational acceleration in the prototype. Thus the response in the model may be different from that in the prototype when gravity forces are important. liowever, when gravity forces are much less than the other forces presented, such as the case of high-velocity package impact, there should not be significant difference between the model response and prototype by not scaling gravity.
Geometric scaling also dictates that fluid viscosity p be smaller (larger) by a factor S in the model than in the prototype. For package application, viscous forces can be neglected. If viscous forces are deemed important and fluid behavior is determiner 8 by density and viscosity alone, this difference can be accounted for by selecting for the model a fluid that has the same density as the fluid in the prototype but a viscosity smaller (larger) by a factor S. l J
!t l Another nonsealing effect is the fracture toughness of the material. Fracture j behavior is a function of absolute size. Fracture behavior of a material depends on the scale size of the structure in two ways. First, fracture depends on the size of the imperfections inherent in the material [Ref. A 9). This dependence is significant only when the size of the imperfections is of the order of magnitude of the material thickness. Thus, for package applications, no differences are expected. Second, it has been observed that the fracture toughness of a materialincreases as the material l thickness decreases below some critical thickness. This effect is the opposite } direction from that indicated by the scaling laws. Thus, the difference in fracture toughness between the model and prototype is usually small unless the scale factor is very large. The last variable the scaling laws requires is the material rate constant t should bc l reduced (increased) by a factor S from the model to the prototype. However, the same materials and therefore the same material rate constants are used in the model j and prototype to preserve complex stress-strain behavio. The effect of this nonscaling variable depends largely on the rate sensitivity of materials used in the construction of the package. The material strain rate effect is most prono' meed in the vicinity of the yield stress for mild steel, while the strength at larger strains t l increases only modestly with increase in strain rate. The grout and some other l proposed energy-absorbing materials' strain rate sensitivity is uncertain. Thus, the l material strain rate effect could become a major contributing factor for the difference l in overall responses between the model and the prototype. Material testing as well as the package model testing in the range of intcrest are required to determine its significance. It should also be pointed out that geometrical scaling is not ensured when tearing, cutting cr ductile brittle transitions occur during a structural response. A.3 References i A-1. W.G. Soper and R.C. Dove, " Similitude in Package Cushioning," J. of Applied i, Mcchanics (June 1962), pp. 263-266. A 2. W.E. Baker,"Modeling of Large Transient Elastic and Plastic Deformations of Structures Subjected to Blast Loading," J. Applied Mcch. Trans. ASME, Vol. 27 (September 1960), pp 521527. A-3. A.A. Ezra and F.A. Penning," Development of Scaling Laws for Explosive Forming," Experirnental Mechanics, Vol. 2 (August 1962), pp. 234 239. A-4. W.G. Soper, "Dynam!c Modeling with Similar Materials /' Shock and Vibration, ASME publication, New York, NY (1963), pp. 51-56. 1 A.S. P.R B. Dallard and J.C. Miles," Design Tools for impact Engineers," Structural Irnpact and Crushworthiness, V01. 2, pp. 369 382. 8 -46 j
i I ~ ? I A-6. N. Jones, " Scaling of inelastic Structures Loaded Dynamically," Structural - Ernpact and Crashworthiness, Vol.1, (1984), pp. 45-74. i [ - A-7.- P. McConnell, "A Feasibility Study for Scaling Ferritic Spent Fuel Casks for. Drop Tests," Proc. PATRAM Conf. (1983), pp. 918-925. i ~ A-8. T.A. Duffey, " Scaling Laws for Fuel Capsules Subjected to Blast, Impact and ; Thermal Loading," Proc. Intersociety Energy Conversion Engineering Conf., . SAE paper No. 719107 (1971), pp. 775-86. r A-9 V. Weiss and S. Yukawa, " Critical Appraisal of Fracture Mechanics," Fracture l l Toughness Testing and its Applications, ASTM Technical Pub. No. 381 (1964). t t i I I t l I i i i i c a I <47-i W W *
- g+ 4mrw er,wy ag, egg wwMpygggr,,g,_
p
4 1 o .~* j APPENDIX B s GROUT DESCRIPTION i 1 i The formulation and mixing procedure for the grout used in the mechanical test i specimens, impact test models, and impact target blocks are given below. They are derived from a procedure for preparing grout used in the transport packaging i for the Shippingport reactor vessel [Ref. B-1), B.1 Formulation Incredients Cement: Portland - Medusa Type 11 (A'STM-C-1:.7 l Fly Ash: Type F (ASTM-C-618) Fluidifier: Interplast-N (CRD-C-619) l Sand: Clean masonry (ASTM-C-33) Sieve passines Mix auantities (% weicht)- No.16 100 % Cement: 22.44 % i No. 30 90 -100 % Fly _ Ash: 22.44 % l No. 50 35 - 95 % Fluidifier: 0.15 % l No.100 5 - 10% - Sand: -37.40 % 4 Water: 17.57 % i l B.2 Mixing Procedure A standard, rotating-drum aggregate mixer is used to mix the grout ingredients. i The mixer is thoroughly cleaned before each use to avoid contamination of the grout. While the mixer is rotating 16 i 5 rpm, sand and 80% of the total allowed water j are added to the mixer according to the specified mix quantitic. -Next, the cement l and fly ash are added according to the specified mix quantities. Next,10% of the l total allowed water is added and the mixer is rotated approximately,50 turns. b The mixer rotation is stopped and the grout mix is allowed to rest for 15 i 2 minutes. ~ The rotation of the mixer is reestablished at 16 i 5 rpm - The fluidifier and the remaining 10% of allowed water are added according to the specified mix quantities. ' Next, the mixer is rotated approximately 100 turns. i 1 The grout mix is poured into prepared casting containers (test specimen tubes, test F models, or target forms). While the grout is curing, the casting containers are I f .48 - . ~.. . ~. ..... ~...
l s' 4 covered with plastic sheet to inhibit rapid evaporation of water from the grout. During curing, the test specimens and test models are stored indoors at room temperature (22 i 2 'C) and the target blocks are stored outdoors at ambient temperature (20 i 7 'C). B.3 Reference i B-1 W.O. Greenhalgh, R.J. Cash, and D.M. Burgess, Shippingport Reactor Vessel / Neutron Shield Tank Package Certification Activity, Benchmark Drop Test Model, Concrete Mix and Fill Procedure, SPC-TP-0002, Rev. O, Westinghouse Hanford Co., Richland, WA (January 1988). y i l a 1 i l 4 9 i
, __-,~..... -.~..--. i t I APPENDIX C i GROUT PROPERTY MEASUREMENTS i l The mechanical properties of grout material used in the package models as the energy absorber and as an impact target to simulate soft rock were measured. i Specimens were prepared following the formulation and mixing procedures outlined in Appendix B. All specimens were 5.08 cm diameter by 10.16 cm long and were cured in plastic mc'3s. Properly cured specimens were sent to a mechanical testing laboratory for unconfined compressive strength measurements and for-l confined property measurements. C1 Grout Sample Unconfined Property Measurements i Table C.1 lists the unconfined grout properties measured and reported by the testing i laboratory, j Table C.I. Unconfined grout property measurements Sample Cure-time Elastic modulus Poisson's Ratio . Compressive ] No. (days) (MPa) - strength (MPa) i 1 14 15,169 0.200 24.00 4 21 15,997 0.246 25.95 1 5 21 16,410 0.228 27.00 } 6 21 16,204 0.242 25.81-7 28 23,443 0.245 30.58-i l C2 Grout Sample Confined Property Measurements { Table C.2 lists the confined grout properties measured and reported by the testing. laboratory. t i s' 4 50 ( _.... -. _,...~ _ _,...,_.,~. _,. _ __ ., _ _.... _... ~.. . ~..
4' l' 3 Table C.2. Confined grout property measurements I Sample. Confining ' Bulk Elastic ' Poisson's _ Compressive j No. Pressure Modulus Modulus Ratio Strength (cure time) - (MPa) (MPa) - (MPa) (MPa) q l (days) l 13 (14) 137.93 14,530.1 3,944.0 0.11 248.0 14 (14) 2 %.90 19,951.0 7,490.0-0.21 327.0 f 17 (28) -34.48 11,991.4 2,080.0 0.00 74.5 l 18 (28) 137.93 _ 11,152.4 3,170.0 0.05; 322.0 l 16 (28) 206.90-20,949.7 11,200.0 - 0.21 352.0 12 (28) 206.90 12,975.7 5,350.0 _0.09 - 377.0 L l. The actual recorded stress-strain and pressure-volumetric strain plots for the above cases _ were all reported by the testing laboratory. Figure C-1.shows a typical stress-i strain plot from a triaxial compression test. Figure C-2 is the hydrostatic pressure-- i volume strain plot for the same specimen, i i i
- -4 c
.,n,--.,v,,
_ = _ _ t ( 180 N A i 'So y ft Q-i = 140 1 l/ 120 -V [ L B 100 E \\ l p 05 60 l E l 4 4 [ q: I N ( j 0 5 -4 -3 2 -1 0 1 2 3 4 5 ]- (Radial) Strains (Axial), % l l Fig. C-1 Typical grout specimen stress-strain plot from a triaxial compression-l test. Cure time = 28 days, confining pressure = 207 MPa (30 ksi). i 250 i i a b .b $ f l f( j )-
- /","'8 887
[ l N / ) ct / 1 [dP dt,=2D9.497 / -o '/ l 50 O O 1 2 3-4 5 6 '7 8-9 Volumetric Strain, %- Fig. C-2 Typical grout specimen hydrostatic pressure-volumetric strain plot. l Cure time = 28 days, confining pressure = 207 MPa (30 ksi). l i 52 - j
I \\ APPENDIX D OTHER ENERGY-ABSORBING MATERIALS Another potentially promising material that could be used as the energy-absorbing material in the package design is high-density foam. Foams usually come in a wide range of densities. Urethane foam of different densities were used in preliminary impact studies. The density varies from 0.32 g/cu em to 1.28 g/cu cm. D.1 Foam Material Properties The relationships between the collapse strengths and other mechanical properties of the foam are discussed in Refs. D-1 to D-3. Foam properties were supplied by a plastics manufacturing company. The compressive strength of foam as well as its elastic modulus in compression varies log-linearly with respect to its density. Figures D-1 and D-2 show these relationships. D.2 Foam Pressure-Volume (P-V) Relationship One of the principal energy-absorption mechanisms in foam is its volumetric compaction, yet no reliable pressure-volumetric strain (P-V) measurement on high density foams can be found in the open literature. In the absence of experimental data, a conservative assumption is made to equate its volumetric compaction strength to the uniaxial compressive strength. A piecewise linear approximation to the foam P-V relationship is adapted. Thus the volume of the foam is assumed to vary linearly with the compressive pressure until its original volume has been compressed down 1 percent. From this point on, the crushing strength of the foam is assumed to remain essentially constant until it has been compressed to about 30 percent of its original volume. The crushing pressure required for further compression begins to increase very rapidly. This phenomenon is called " lock-up," and is a common phenomenon among most porous or crushable materials. As an example, to calculate the mechanical properties for the 0.96 g/cu cm foam, Fig. D-1 is used to estimate its uniaxial compression strength, which is about 50.3 MPa. This corresponds to a crush pressure of about 16.8 MPa. The slightly higher crush pressure was selected to ensure that deviatoric yielding occurred before volumetric yielding takes place. D.3 Package Finite Element Model and Related Impact Analysis The DYNA 2D finite element computer code was employed in a study of the package impact analysis. Figure D-3 shows the finite element mesh of the three-layer reference package model generated for the analysis. The inner containment is an oval cylinder,15.24 cm diameter by 29.5 cm long, and is assumed to be essentially a 4 \\ rigid body with no permanent deformation after an impact. The foam energy-absorbing material surrounding the containment is 63.5 cm diameter by 111.8 cm long. The outer wrapper is made of a ductile steel such as stainless steel type 304 with a thickness of 1.02 cm. D.4 DYNA Foam Material Model Representation Because of the lack of more accurate foam material property data, the simple ' soil and crushable foam' material model in the DYNA code is chosen. Again, due to the lack of actual foam material deviatoric yield stress dependence on pressure data, both a and a are assumed to be zero,i.e., deviatoric yield stress is independent of 3 2 the pressure amount. Table D.1 summarizes the material constants used in all DYNA simulations using foam as the energy absorbing materials. Table D.I. Foam material property used in DYNA finite element model Density 0.32 0.48 0.64 0.96 1.28 { (g/cm3) Young's mod., 137.9 268.9 413.7 723.9 1,137.7 E, (MPa) Poisson's ratio 0.4 0.4 0.4 0.4 0.4 Shear mod. G 49.25 96.04 147.75 258.57 406.32 (MPa) Bulk mod., K 229.84 448.18 689.51 1,206.65 1,896.16 (MPa) Uniaxial comp. 8.96 14.48 24.13 50.33 89.64 strength, (MPa) a0 3,885.4 10,135.8 28,152.8 122,457.4 388,402.4 al, a2 0 0 0 0 0 P-V relationship In (V/Vo) 0 0 0 0 0 0 -1.005E-2 3.0 4.8 8.1 16.8 29.9 -1.204E00 3.0 4.8 8.1 16.8 29.9 -2.303E00 3,000.0 4,826.6 805.4 1,678.9 2,989.0 -4.605E00 4,378.4 6,205.6 2,184.4 4,437.0 4,368.1 ~
( D.5 Analysis Results Quite a few cases of package impact analyses onto an unyielding surface were made using the ' reference' model geometry at different package impact velocities. The analysis was also extended where the size of the energy-absorbing foam material is one and a half times the reference size while maintaining both the inner containment and the outer wrapper at the same dimensions and thickness, respectively. Tables D.2 and D.3 list the results for the reference size and the enlarged size. Since the foam's crushing strength varies linearly with its density, when the package containment peak g value is plotted versus foam density / crushing strength for a given package impact velocity,it should be reasonable to expect that the shape of the curve will exhibit a concave upward characteristics. This is because at relatively low foam-crushing strength, there is not enough " cushioning" provided by the foam material. As a result, the package containment sees a higher deceleration value. As the crushing strength increases, containment deceleration level drops down rapidly. However, at relatively high foam-crushing strength, the foam becomes too " stiff" for its intended purpose; thus the curve starts to rise again. Therefore, for a given package impact velocity, we can expect the containment peak g versus foam-crushing strength curve to be of a concave upward shape, passing through an optimal foam density / crushing strength point for that given package configuration as well as package impact velocity. This is indeed the case. Figures D-4 and D-5 plotted the relationship between the package containment peak g level versus foam density for two different package impact velocities at 91.44 m/s and 137.16 m/s from the analytical results presented in Tables D.2 and D.3. D.6 Foam Energy Method Representation An interesting relationship emerges when one forms an " energy-absorption" index, and plots the package containment peak g vs. the index. The index is defined as the kinetic energy of the package containment per unit volume of available energy-absorbing material, or mv2/2 / V where m is the mass of the containment, v is the package impact velocity, and V is the volume of the "available" energy-absorbing material. If the foam volume directly undr.rneath the containment is used to approximate the amount of "availabic" energy-absorbing material volume, and the index value is i calculated for all the cases analyzed and listed in Tables D.2 and D.3, a relationship as shown in Figure D-6 is obtained. This figure shows that an essentially linear 4 Table D.2. Package impact on unylelding surface. Package dimension is 25 cm diameter x 44 cm long x 0.4 cm thick. Foam Outer Shell Density Impact Impact Contain. Maximum (g/cm3) Velocity (m/s) Time (ms) Peak G Plastic Strain (%) 032 91.44 33 2,510 30.0 0.48 91.44 3.2 3,460 30.0 O.64 91.44 3.0 4638 32.0' O.86 91.44 2.8 7,624 31.0 1.28 91.44 2.4 9,255 29.0 032 137.16 3.5 5,318 63.0 0.48 137.16 3.5 4,591 61.0 0.64 137.16 3.0 5,738 59.0 0.96 137.16 2.5 9,193 54.0 1.28 137.16 23 13,618 42.0 1.28 182.88 23 14,115 61.0 Table DJ. Package impact on unyielding surface. Package dimension is 37.5 cm - diameter x 66 cm long x 0.4 cm thick. Foam Outer Shell Density Impact Impact Contain. Maximum (g/cm3) Velocity (m/s) Time (ms) Peak G Plastic Strain (%) 0.32 91.44 43 2,135 36.0 0.48 91.44 43 3,075 34.0 0.64 91.44 4.2 4,749 32.0 0.96 91.44 3.9 7,306 35.0 1.28 91.44 33 8,608 33.0 032 137.16 4.4 3,331 61.0 0.48 137.16 4.2 4,229 58.0 0.64 137.16 4.0 5,455 58.0 0.96 137.16 3.5 8,926 56.0 1.28 137.16 33 9,868 38.0' relationship existed between the containment peak deceleration level and the " energy-absorption" index. This relationship is useful in the preliminary sizing of the package dimensions for high-velocity impact studies. -56
4-l. L 4 D.7 References 1 .s D-1. S.K. Maiti, L.J. Gibson, and M.F. Ashby, " Deformation and Energy Absorption Diagrams for Cellular Solids," Acta Metal., Vol. 32, no.11, pp.1963-1975 (1984). I-D-2. L.J. Gibson, et al., 'The Mechanics of Two-Dimensional Cellular Materials," i Proc. R. Soc. Land., A328, 25-42 (1982).- D-3. L. J, Gibson and M.F. Ashby, "The Mechanics of Three-Dimensional Cellular Materials," Proc. R. Soc. Lond., A328,43-59 (1982). ~ 1 l i i i I i-1' l ) i e i ~. ~. - = - J A k i t b 1 d. % I N 5 ....i .r.-
- "*~~""'~~
4so.ee se e==., e ,l wem.se.C+ee. se.a.m i se LA41.e40aA** De4ses 4 m gais / l eased T.. S.- ) = i, a 7 k e p-j i 1 .4 I 4 / - 1: --y - -p p w i 1 j e 1. f ~~ / a / 4 f o a eu. e e ...e j e tff 't t t f f f fI t o,,i,. I f f Fig. D-1 Foam compressive strength vs. density relationship 3 .e.. i. u .... n. u. .u .. i -1 ).
- m..se.
.. a s
- ' -*=
444 de 400 k. .so eae v. t i ~g 4, e 838~ 10 000 4> .e e. e oog
== ' 4 000 =-- k: 4l 4 37,. I. i 5 3,co y,au twswa a y i. h iu W n 900 ~
- -=='-
.8'
- 4. 8.
e egg
- =-
- = =
- 37.
400 4.34 300 l, i c -... i. 3. 4 DID4tif. 4/na Fig. D-2 Foam elastic modulus vs. density relationship a .[ ~ . o +, - .m.. s v y
1 .g 1, - i d f 1 ' / l l / p N l 1-^ l s i [ i a -i-n a" 4 4 y i ' ~ s, ~~ s s q s s i f. Nr' 1-s i i-i- Fig. D-3 Finite element mesh for the package analyzed i f i k i j .a og -20000 i .W 4 e8 15000 e i C-i 300 fps ' E 10000 - -450 fps j m .a - 3 5000 - m U-o- 0 O. 20 40 60 80L -100 Foam density:. (Ib/cu.ft) i Fig. D-4 Peak containment vessel G vs. foam density plot 4 1, 1
. ~.. -. .j - I' 1-i S i 10000 j te i x 8000 - a o 300 fps l 6000 - 450 fps j 6 .6 4000 - 2 c O O 2000 0 20 40 60 80 100 Foam density (Ib/cu.ft) ~ Fig. D-5 Package containment vessel peak G vs. foam density plot t i. d
- 1. -
16000 l M~ o 14000, i 3 c5 12000 - 8 o m-30#. 10000 40#. }. E a U 8000 - a 60# l 6 e-80# .6 6000 2 .c.s. s i 4000 - U 2000 i ' O 1000 2000 3000 4000 5000: 6000 -my2/2[V Fig. D-6 Peak containment vessel G vs. ' energy-absorption' index for foam 1 material i ' 60. .-}}