TRABSIENT - PDF document

1 DIRECT TRABSIENT ANALYSIS OF A PUZE
DIRECT TRABSIENT ANALYSIS OF A PUZE ASSEMBLY BY AXIS~RIC SOLID ELEWENTS Chung C. Dai and 6. S. Yang* Advanced Technology & Research. Inc. = 20866 John Titus Barry Diamond Laboratory Adelphi. 1ID 20783 i.e., nose. collar and sleeve. was designed to survive severe maxi~um base acceleration of 20.000 G. Experiments shoved that hoop failure occurred in the ccllar after the impact. They

2 also shoved that by bonding the coll
also shoved that by bonding the collar was able to survive the same impact. To find out the effectiveness of the bonding quantitatively. axismetric solid elements TRAPAX and TRIAAX vere used in modelling the fuze and direct transient analysis was performed. The dynamic stresses in selected elements on the bonded and collars vere compared. The peak hoop stresses in the unbonded collar vere foun

3 d to be up to three times higher than
d to be up to three times higher than those in the bonded collar. IUSTBLLA failure in mbonded value runs aodels prior KO the transient runs. Their experiments. Coarents on the use of the I4PCAX cards. the existence and contributors of the several nearly H732 E2 Artillery Proximity Fuze corprises an aluminum fuze body vhich houses a safety and arming module. detonator bulkhead, and pover supp

4 ly aft of a center bulkead. Forward of
ly aft of a center bulkead. Forward of fuze processor (PPO) nose cone vhich is secured to the fuze body vith a 30% glass reinforced PPO retaining collar. - When a projectile is fired frm a gun. the fuze is subjected to 20.000 G's. In the laboratory. a machine -Jas developed to simulate this loading. The fuze is secured to a rail guided eteel fiztcrc vtich is struck from the side a steel weight.

5 The impact is tailcztd co give a 2
The impact is tailcztd co give a 20.000 G peak lateral ..zc;ge. temperature experienced Ca.-ures -&en temperature conditioned -40J, :ailed vithout exception. The major failure mode, causing a frcm end to e~3. See Figure 3. It was discovered that by solvent bonding the retaining collar to the nose cone during assembly. the low temperature failures could be avoided. Since the mechanism by vh

6 ich the solvent bond alleviated the pr
ich the solvent bond alleviated the problem was not well understood nor was the safety factor of the bond lmovn. it was feared that future modifications to the fuze could not be evaluated vith respect to the impact on the fuze joint integrity. Therefore. HASTRAN vas used to wdel the forward section of the fuze in an att-pt to better understand the behavior of the current design. ana better predict the beh

7 avior of future designs. The given sec
avior of future designs. The given section consists i.e., nose. collar and sleeve. As depicted in Figures 4a. 4b. and 4c. 105 WAX and 9 TBW elements were used to model this fuze and it* interior electronic parts. These elments eigen value perfcrmed pricr to the final transient runs. Wlnerical results fra the static and eigen the available Constraint conditions WDEI. I A cantilevered. ho

8 llow slender cylkder. as depicted in F
llow slender cylkder. as depicted in Figure 5a. was modelled by TRAPAX eleaents as a preliminary study. Stress analysis of this cylinder was performed. Axial presscre load or concentrated transverse tip loading was separately applied. Merical results from these test cases provided basic understanding on TRAPAX'S behavior and characteristics. Axial pressure loading condition was first considered. see Figu

9 re 5b. As expecred. HASTRAN results
re 5b. As expecred. HASTRAN results cylinder's cross sections vere di~?lace4 uniformly and axial stress was developed uniformly across each cross section. Figure 5c shows the trarsverse tip loading condition. The cylinder was modelled by regular 8 x 2. 16 x 2. 32 x 1. 32 x 2. 64 x 1 meshes. Both the tip deflection and the bending stress at point A were checked for convergence vith respect to mesh ref

10 inement. As noted '.n with the same nu
inement. As noted '.n with the same number of elements. 64 x 1 mesh gave better resuls than 32 x 2 mesh. Blsed on this observation, we used only one layer across the thickness of the real model. TEST HODEL I1 slender cylinder 6) vas used in the remaining preliminary studies. It was modelled also by TRAPAX eiements but with 4 x 1 or 5 x 1 mesh. This model served for testing MASTUN rigid for

11 mats 3. 9, and 12. simulating the grese
mats 3. 9, and 12. simulating the gresecribed also for NASTRAN checkpoint and restart features. Free vibration analysls was performed on the cantilevered lo?low cylinder in Figure 6a. Different frequency ranges were specified requesting the extraction of natural frequencies and mades, Given V lc 30 ftlsec after the impact and the maximum acceleration of 20,C)OOg. depends on assum tion of its s

12 hape. PI For a 8in~t shaped impulse.
hape. PI For a 8in~t shaped impulse. aboot 70 x 10 sec. For a sin'wt about 90 10 sec. Impulses of both shapes were tested on We1 11. Howver, only the sinAot shaped impulse was employed in the analysis of the real model because it sirrulter the real impulse more closely. Several schemes to input the prescribed base acceleration a(t) were also carefully examined. First. a nearly rigid lay

13 er with huge mass X was attached to th
er with huge mass X was attached to the bottom the original F(t) = Xa(t) was to this vays to achieve the desired prescribed base motion. The dynamic force can be directly applied to the COSINE term of the ring lo. 2 in Figure 6b. It can also be applied, using POINTAX cards. at selected points in the bottom approaches gave "WINTAX" approach was taken on the real model. The appropriate Elp r

14 atio was determined to be of the order o
atio was determined to be of the order of 10- . This specific ratio gave an uniform base acceleration as desired. In addtition. both COW and TRAPM elements were tested out IS the base layer. The TWAX elesent was employed in the real model. Modal transient analysis was performed on this test model. r -ever. as noted from the real raoge. short that extremely high without using vhich in turn will

15 make it extremely time-consuming. Theref
make it extremely time-consuming. Therefore. direct transient approach was taken for the analysis of real model. STATIC AYALYSIS bonded and unbonded For loading case. nodes 17. listed fcr of harmonic 2) Based on its converaence characteristics. AXIC = the transient different parts lxamined, and the MPCAX cards were modifid simple experiments load onto Loads were incrementing performed for bond

16 ed and experiment Note that the noticeab
ed and experiment Note that the noticeable differences say have been the force gage made a slight indentation EIGEN Eigen value analysis vere performod to further check the appropriateness of the NASm model. WPCAX condition; between nose-collar-plate repeatedly examined and these conditions arrangement induces only relatively while maintaining stability of the the seventh WCAX number should he form of

17 (Integer 20) instead of (Integer 70).
(Integer 20) instead of (Integer 70). modes may the model. fundamental frequencies modes were both bonded are very close together identifed, very interestingly. that harmonic contributed to one DIRECT TRANSIENT ANALPSIS -4 Transient responses from t = G sec to t = 270 x 10 sec, times the woulse both bonded and unbonded time steps listed in at = 2 x lo-' sec was chosen for the time history co

18 mpared very well with plan? and 105 an
mpared very well with plan? and 105 and compared between both the hoop peak hoop stress ic unbonded collar up to three times the bonded This explains the hoop failure in the unbonded collar was observed from the axial time histories axial stress significantly higher 11 lead to following conclusion: By bonding the and the collar, the nose take dynamic will thusly factor of the hoop effectively avoided in bond

19 ed collar. SOLUTION CONVERGENCE WITH RE
ed collar. SOLUTION CONVERGENCE WITH RESPECT MESH REFINEMENT --- x 2 1 32 x 2 64 x 1 ANALYTICAL ASPECT RATIO 7.5 3.75 0.94 1.875 0.47 ---- CONVERGENCE OF UlTH RESPECT TO THE NUMBER OF HARMONICS (BONDED COLLAR) TABLE 3. COMPARISON OF STATIC RESULTS (TmSVERSE LOAD AT 0.27" FROM NOSE TIP) NO. OF HARMONICS 3 4 7 1 LATERAL DEFLECTION NODE 9 -0.00676 -0.00737 -0.00767 NASTFUN -0.00676 -0.0082 1 -I Y

20 CASE i3Oh'DED UNBOLXIED EXPERIMENT -
CASE i3Oh'DED UNBOLXIED EXPERIMENT -0.014 -0.020 NODE 17 -0.OG467 -0.00467 NODE 21 -0.00375 -0.00375 -0.00467 ! -0.00375 CONSTRAINT CONDITIONS CCLm AID PLATE C 3 7 4 6 L 6 0 237 240 243 246 249 252 255 258 260 263 364 365 370 UNBONCZD z 0- 3 G- 3 0-3 0-3 0- 3 0- 3 0- 3 0-3 0-3 0-3 0 0.1 0.1 r 0-3 0- 3 0-3 0-3 0-3 0-3 0-3 0-3 0-3 0-3 0-3 --- 0.1 --- r 0-3 0-3 0-3 0- 3 0-3 0- 3 0-3 0-

21 3 0-3 0-3 0-3 ! --- 0.1 --- 8 0- 3 0
3 0-3 0-3 0-3 ! --- 0.1 --- 8 0- 3 0-3 0-3 0-3 0-3 0-3 0- 3 C-3 0-3 0-3 0-3 --- --- --- 8 0- 3 0- 3 0- 3 z 0- 3 I 0-3 I 0- 3 0- 3 I --- I 1 0-3 I --- I 0-3 i --- i 0-3 I --- ' 0 I --- I Ovl I --- 0.1 EIGEN VALUE ANALYSTS BONDED NOSECONE WDE EXTPACTION NO. CRDEP EIGENVALUE EIGEN VALVE ANALYSIS UNi3ONDED NOSECONE lWDE EXTRACTION NO. ORDER IAN CYCLIC FREQUENCY FREQUE

22 NCY REAL EIGEN RAP1 AN 3 EOU TEN NA
NCY REAL EIGEN RAP1 AN 3 EOU TEN NATURAL FUZE EXTRACTED TABLE 6. CONVERGENCE OF TRAKS?EK SOLUTIOKS WITH RESPECT TO TIME STEP SIZE - I riEMOTIOK rt=.p~ec 1 Acceleration At: I at-1~~;~ .I~=?PS~C At=o.5b'S~C' 5-92 x 105 j t = !UU sec --- 0.22 x lo5 I 9.13 x lo5 9.09 x :J~ 1 I , Displacement At: t = 8 sec t = IOU sec I I ! I 6.08 x lo5 5.98 x 105 I t=8r sec 6.7 x