Pr 32.413/3:1868
CONFIDENTIAL DECLASSIFIED
NATIONAL DEFENSE RESEARCH COMMITTEE ARMOR AND ORDNANCE REPORT NO. A-218 (OSRD NO. 1868)
DIVISION 2
THIRD PROGRESS REPORT
ON
PLASTIC DEFORMATION OF STEEL UNDER HIGH PRESSURE
by
P. W. Bridgman
This document contains information affecting the national defense of the United States within the meaning of the Espionage Act, U.S.C., 50, 31 and 32. Its transmission of the revelation of its contents in any manner to any unauthorized person is prohibited by law.
LIBRARY
UNIVERSITY OF WASHINGTON
NOV 14 1950
Copy No. 87
CONFIDENTIALARMOR
NATIONAL DEFENSE RESEARCH COMMITTEE AND ORDNANCE REPORT NO. A-218 (OSRD NO. 1868)
DIVISION 2
THIRD PROGRESS REPORT
ON
PLASTIC DEFORMATION OF STEEL UNDER HIGH PRESSURE
by
P. V . Bridgman
Approved on September 23, 19R3
for submission to the Division Chief
A IV gDjz


P. V:. Bridgman, Consultant
Division 2
Merit P. White, Secretary
Division 2
Approved on September 23, 1Rh3 for submission to the Committee
GOMTDENT1AL
^2J '^'7 Zp1	£ c ‘ c.
- John E. Burchard, Chief
Division 2
Structural Defense and OffenrPreface
The work described in this report is pertinent to the project designated by the Navy Department Liaison Officer as NO-11 and to Division 2 project P2-3O2.
This work was carried out and reported by Harvard University as part of its performance under Contract 0EKsr-201.
Initial- distribution of copies of the report
Nos. 1 to 25, inclusive, to the Office of the Secretary of the Committee for distribution in the usual manner;
No. 26 to R. C. Tolman, Vice Chairman3 NDRC;
No. 27 to R. Adams, Member, NDRC
No. 28 to F. B. Jewett, Member, NDRC;
Io. 29 to J. E. Burchard, Chief, Division 2;
No. 30 to W. Bleakney, Deputy Chief, Division 2;
No. 31 to O’. F. Davidson, Office of the Chairman, NDRC;
No. 32 to R. A. Beth, member, Division 2;
No. 33 to H. L. Bowman, Member, Division 2;
No. 3h to C. W. Curtis, Member, Division 2;
No. 35 to C. W. Lampson, Member, Division 2;
No. 36 to W. E. Lawson, Member, Division 2;
No. 37 to H. P. Robertson, Mission London;
No. 38 to F. Seitz, -’ember, Division 2;
No. 39 to A. H. Taub, Member, Division 2;
No. l>0 to E. 3. Mils on, Jr. Member, Division 2;
Nos. hl and h2 to R. J. Slutz, Technical Aide, Division 2;
?To. h3 to Army Air Forces (Brig. Gen. B. W. Chidlaw);
Nos. hh and h5 to Corns of Engineers (Col. J. H. Stratton, Lt. Col. F. S. Besson, Jr.);
No. h6 to Ordnance Department (Col. S. B. Ritchie);
No. h7 to H. P. White, Technical Aide, Division 2;
No. h8 to Corps of Engineers (Lt. Col. S. 3. Smith);
No, h9 to Watertown Arsenal (Col. H. H. Zornig);
No. pO to Aberdeen Proving Ground (0. Veblen);Nos. 3l and 32 to Bureau of Ordnance (Lt. Comdr. T. J. Flynn, A. We rtheimer) ;
No. 33 to U. S. Naval Proving Ground (Lt. Comdr. R. A. Savyer);
No. 33 to David Taylor Hodel Basin (Capt. W. P. Roop); No. 33 to Bureau of Ships (Lt. Comdr. R. VI. Goranson); No.. 36 to Bureau of Yards and Docks (War Plans Officer); No. 37 to U.S. Naval Research Laboratory (R. Gunn);
No. 38 to D. S. Clark5 Consultant, Division 2;
No. 39 to A. Nadal, Consultant3 Division 2;
No. 60 t ' P. VI. Bridgman, Consultant, Division 2.
»
The NDRC technical reports section for armor and ordnance edited this report and prepared it for duplication.contents
Pag
-Abstract
Section
Introduction ............................
•
Distribution of stress across the nocked specimen ...............................
Relation between flow stress and natural strain .................................
Detailed results and discussion .........
Conclusion ............
20
List of Figure
Fi
igurc
7.
Two curves relating to the effects of necking in tension specimens .................. h
Corrected true stress at fracture and maximum pressure during pulling versus natural strain for plate 18A............. 1 £
Ratio of area of tensile fracture to area of neck versus maximum pressure during pulling for plate 18A ............... 1 $
Corrected true stress at fracture and maximum pressure during pulling versus natural strain for plate 13-7................ 16
Ratio of area of tensile fracture to area of neck versus maximum pressure during pulling for plate 137................... 16
Corrected true stress at fracture end maximum pressure during pulling versus natural-strain for plate 17?......... 17
Ratio of area of tensile fracture to area of neck versus maximum pressure during pulling f or plate 177........................... 17Figure
11 ,12.
13.
Ih.
19.
16-19.
Corrected true stress at fracture and maximum pressure during pulling versus natural strain for plate 6X1 ...................... 18
Ratio of area of tensile fracture to area of neck versus maximum pressure during pulling for plate 6X1 .......................... * 1 8
Double neck on a tension specimen from
plate 17F .................................. 19
Photograph that shows the method of sectioning a tension specimen ..................... 21
Specimens from plate Clli broken in tension under atmospheric pressure; and under hydrostatic pressures of IjOOO, 10,100 end 1 9,800 kg/cm2 ......................... 21,23THIRD PROGRESS REPORT ON
PLASTIC DEFORMATION OF STEEL UNDER HIGH PRESSURE
Abstract
Tests of the mechanical properties of steels under high hydrostatic pressure which were described in previous reportsl^/were continued, using four samples' of armor plate of varying ballistic qualities. The conclusion, in accord with the results of other lines of investigation, is that ballistic behavior is closely associated with' fairly obvious characteristics —- such as inhomogeneity, brittleness, and so forth — which can be investigated by standard methods. The conclusions regarding strain hardening and conditions for fracture, given in the earlier reports, have been improved, but not essentially changed, by considering the true stress state in the neck of a tensile specimen,
1
Introduction
The present work constitutes an attempt to. answer one of the
original questions that was deemed important when this.program was
initiated, namely, whether by taking account of the properties of
plate under pressure it might be possible to anticipate ballistic
failure. The question was of interest because at that time'thebe
were many cases of obscure correlation between ballistic failure
and the more usual physical properties? but It could not be-'
1 2/
answered in the two earlier reports—~ because samples of plate-which had.failed to meet the ballistic tests were not then available.
In this:report data are given for the effect of hydrostatic pressure on the tensile properties of four samples of-armor plate.
1/ P. W. Bridgman, Plastic deformation of steel under: high pressure, NDRG Report A-?5’ (OSRD No. 91R).	''
2/ P. rt. Bridgman, Second progress report on plastic tion of steel under high pressure, NDRG Report A-162 (OSRD
deforma-
No.2
•H
-P
O
a?
p
•H
rd
o
-p
p
P
P
o
o	
•H "P P 0 P P 0 P w p 0 6 •H O 0 P w 44 £	
	0
bfl	-P
P	cd
•H	r—1
i—1	Ph
H O	<P
P	c
<H	0
O	o
P	p P
o •H	o
-P	-p
O 0	Pl
P	p
•H	r~)
dJ	P
O	O •H
-p-	d4
i—1	P 0
0	Ph
H	P
H	0
P	Ph
P P	W
Ph	P
P	0 g
0	•H
44	O
P	0
-P	Ph
m P	co (0
0	co
.§	0
•ri O	
0	o
Ph	•H
ro	rP
44 -P	■P
bfl	b.0
P	
0	•H
hP	r—1
p
I
o
0
P
•H
dJ
m
V)
0
5
o
•H
44
-P
P
•H
P
0
P
-P
Pl
o
-p
o
p
bO
p
•H
H
H
O
P
CH
O
P
o
•H
-P
O
0
P
•H
Td
O
-P
P
P
rH
g
£
P
0
Ph
P
0
Ph
p
0
44
P
-P
P
0
.§
o
0
Ph
w
<p
p
co
•H
K
p
b.0
p
o
HP
f
<
«
<+H
c- 3 -
Partial data have already been presented for one of these plates— but three of the plates are new. The particular significance of these three new plates is that other of their physical properties have already been examined elsewhere. Two of these plates had failed in the ballistic test.
The three new plates were supplied by pr. M. Gensamer, of the Carnegie Institute of Technology, who provided a description of their physical properties^ These are presented in Table I.
The measurements of the present report differ from those of the two earlier reports in that the effect pf orientation in the plates was more carefully examined, tests being made on specimens cut in three mutually perpendicular directions: parallel to face and parallel to rolling (X-direction); parallel to face and perpen dicular to rolling (Y-direction); and perpendicular to face (Z-direction). The fourth plate, designated 6x1 in the first report, was examined with respect to the strain-hardening curve and the effect of pressure on the character of the fracture, points that had received only inadequate consideration in the first report. The particular plate 6X1 was selected because previous experience had shown that it was of good homogeneity.
2• Distribution of stress across the necked specimen
The results presented here have been improved in one important respect in comparison with those of the two previous reports. It is known that the tension is not uniformly distributed across the neck of a tension specimen, and that this lack of uniformity will affect the value of the flow and rupture stresses calculated from tension experiments. Previously, there has been no satisfactory method of calculating the precise distribution of stress across the necked specimen, and in the absence of such precise knowledge it has been the custom to report the results in terms
3/ Reference 1.
U/ In two personal letters dated Dec. 23, 19R2 and Aug. 3,
19h3 .Fig.I. TWO CURVES RELATING TO THE EFFECTS OF NECKING IN TENSION SPECIMENS.•• -•£ -
of the average tension across the neck,._Under ordinary circum-
stances it was to be anticipated that .the correction for lack of uniformity was not very serious; but with the abnormally large degree of necking which prevails in these experiments because of the great increase of ductility under hydrostatic pressure, the correction may be expected to become more important. Fortunately, since the publication of the first two reports, a method has been found for obtaining the precise distribution of stress across the
neck and for evaluating the correction arising from lack of 3/
uniformity—7 The results given in the present report are corrected for this effect; in order to permit comparison with previous work, the uncorrected results are also given. In some cases, where the necking is extreme, the corrected stress of flow may be 7£ percent of the uncorrected stress. So much of the results of the analysis as is necessary for present purposes is reproduced in the two curves of Fig. 1. One curve shows the correction factor in terms of the "natural” strain at the neck, that is, loge Ao/A, where Ao is the initial cross-sectional area and A is the final cross-sectional area; the other curve shows the factor by.which may be calculated the hydrostatic tension on the axis generated by the necking. The detailed analysis evaluates these two factors in terms of the ratio of the radius of curvature of the contour at the neck to the radius of the neck, and not in terms of the natural strain at-the neck. However, it is an empirical result, valid over a wide range of conditions, that to the degree of approximation of the present experiments the ratio of the two radii may be taken to be a universal function of the natural strain. By means of the curves of Fig. 1, the stresses given in the-, two first reports may be corrected. The effect of the correction is to diminish the tabulated flow and fracture stresses; the fraction by which these stresses are to be diminished is greater, the greater the reduction of area.
5/ The detailed analysis is in course of publication by the American Society for Metals.- 6 -
3. Relation between flow stress and natural strain
Application of the correction for necking does not alter the important result of the second report, that is, the true flow stress is a linear function of the natural strain. However, recog nition of the existence of the correction alters the qualitative nature of the picture with regard to fracture. To show what the character of this change is it will be necessary to describe in further detail the results of the analysis.
Consider a tension specimen pulled under normal conditions in a testing machine at atmospheric pressure to such an amount that there is appreciable necking. At the neck there is a distribution of stress which may be analyzed into the sum of two distributions . The first is a tensile stress along the axis of the specimen and constant across the section of the neck; the second is a hydrostatic tension which vanishes at the outer surface of the neck and which increases bo a maximum value on the axis accord ing to a law that is given explicitly in the afore-mentioned paper. On the axis the hydrostatic tension is the product of l°Se 0 + s' r) the constant tension, where a is the radius of the neck and R is the radius of curvature of the contour at the neck. If we represent the stresses in conventional cylindrical coordinates — £ along the axis, r along the radius and © in the plane through the radius at right angles to the z-axis — then the principal stresses are as follows;
At the outside surface	On the axis
rr = 0	rr = F	loge(1	+	1 a\' 2 R
©9 = 0	99 = F	logeO	+	* R?
		logs(1		1	a\ 2	jj) +
zz = F	zz = F		+	
where F is determined by the condition that the integral of zz across the section shall equal the applied tensile load.
In the region of plastic flow before rupture, the specimen is flowing uniformly and the strain is constant across the narrowest- 7 -
section of the neck. The rate of flow or the amount of strain hardening is assumed in the analysis to oe independent of the hydrostatic component of the total stress, an assumption which is amply justified by experiment. Hence a strain hardening curve is to be constructed by plotting the simple tensile component F against the natural strain at the neck>. and this is also the strain-hardening curve if the total stress system is the sum of F and a hydrostatic term. It Is shown in the paper mentioned that F is obtained by dividing the conventional average tensile stress — that is, the total tensile load divided by the cross-sectional area at the neck — by (1 + 2 ~) loge (1 + l|p-. The reciprocal of this expression is the "correction factor" shown in Fig. 1.
Fracture} unlike plastic flow before fracture, does not occur uniformly across the neck but is initiated at a particular point' on tru axis,, where the total stress system has a uniquely determined value. How it is well established by experiment that fracture, unlike plastic flow,, is strongly dependent on; the hydrostatic component of the total stress system, the strain at fracture increasing greatly with superposed hydrostatic pressure. In order to cnaracterize completely the conditions of fracture in a tensile test,, three parameters should be given: the strain, the simple tensile component,, F, and the hydrostatic component, F log (1 +
—	. -	3	bQ \	p /
This complete.specification■could not'be''given before the correction for necking was known.
Ii the tensile test is made in a medium under hydrostatic pressure, then the complete stress system is the sum of three systems, the two just considered and the imposed hydrostatic pressure, inis imposed hydrostatic pressure does not affect the flow parameters, and the strain hardening curve is still to be specified by giving F as a function of the strain at the neck.
The hydrostatic component of the stress at fracture is now, however, the hydrostatic tension arising from the necking, which is equal to Floge(1 +	minus the Imposed hydrostatic pressure.- 8 -
We pass now to a consideration of the actual measurements.
The detailed arrangements of the experiments of this third report}	<
including the dimensions of the specimens,, were exactly like those
in the second reporty^
Li. Detailed results and discussion
The results are summarized in Tables II to and Figs. 2 to 1b. As in the previous report, each entry or point in a diagram is a condensation of an entire curve3 consisting of 10 to 20 readings of tensile load,, hydrostatic pressure and extension. It is not necessary to reproduce the original curves in detail; the significant parameters of these curves are the data given here.
Most of the entries in the tables are self-explanatory. The "Corrected true stress at fracture" of column 9 is the same as the F of Sec. 3. The "Hydrostatic tension on axis arising from necking" of column 10 is Flog8(1 + i I?) • The "Net hydrostatic tension
h,
on axis" of column 11 is column 10 minus the maximum corresponding
pressure of column 2-. The "Net tension at fracture" of column 12	i
is column 9 plus column 11. The values listed in the tables for atmospheric pressure agree fairly 'well with the value reported by Gensamer, obtained under different experimental conditions.
(a)	Relation between true stress at fracture and natural strain at the neck. — These new results in the first place substantiate results previously found. Consider, for example3 Figs. 2, 3^ 8 and 11. In these the true stress at fracture,, corrected for the- nonuniformity of the stress distrioution at the neck, appears to be a linear function of the natural strain at the neck at fracture. In the previous report the uncorrected true stress was also found to be such a linear function. All the test pieces of the present report were pulled to fracture; in previous work the tests were often not carried to fracture, and it was found that the stress-strain relation below fracture follows the same linear relation as that which represents the fractures. In other words, the corrected true stress at the neck, plotted against
6/ Reference 2.- 9 -
the running natural strain at the neck, moves along the straight lines of the figures as strain increases during the course of an experiment. Fracture is a catastrophe that suddenly terminates the experiment without the running point leaving the line or without any other warning. This means that the lines of Figs. 2,
8 and 11 are essentially concerned with the phenomena of plastic flow and strain-hardening and not with fracture. This is emphasized further by reference to the effect of orientation.
(b)	Dependence of strain-hardening and fracture phenomena on orientation. — In Figs, 2, > and 8, the points for the three orientations all lie on the same lines independent of the orientation, which means that for a given natural strain the true stress and therefore the strain-hardening is independent of orientation. However, specimens with different orientations vary in their fracture phenomena. For example, the X-direction of plate 13F breaks under atmospheric pressure at a corrected true stress of 15,300 kg/cm2, whereas the corresponding fracture stress for the Z-direction is only 10,600 kg/cm2. Although fracture in the Z-direction occurs before fracture in the X-direction, up to the instant of fracture the stress-strain curves for the two directions have been identical. The natural strain at fracture for the X-direction is 0.936, and that for the Z-direction much less,
only 0.275.
(c)	Numerical parameters of the strain-hardening curves. —
The numerical parameters of the strain-hardening curves of the four different plates differ only slightly, Vye may take for these parameters (i) the value of the flow stress at a natural strain
of 0.1, which is approximately the point where necking begins in a tension specimen, and (ii) the stress at a natural strain of 3.0. The latter parameter is probably the more significant of the two under the conditions of armor penetration because it indicates the resistance at high degrees of distortion. The stress at a natural strain of 0.1 for the plates 18A, 13F, 17F and 6x1 is, respectively, 12,500, 10,000, 10,h00 and 7,000 kg/cm2, and theTable II. Sunmary of tension tests under pressure on plate 18A
	ns 1 o	JioeN jo is©jy re^oj oj ejn;oBijj ©xfsusj, jo eejy jo on*})		o ns vs o	Os Os	Os O	1	Os -3 O	sO 8	vs -3 Vs O	co -3				ns 1 ■3 o	f*- f*- CM 6	vO r- CSI	CO Os	vs vs	§-	0.382	I*- 8	r* 8				
	CM														CM												
	C		(eTO/23i)	g	R	8	8	8	g	8	8	8			S3	8	g	8	g	g	8	8	8	a	8	8	s
	B		amq.0BJj	SO	sO	co	8	sO	o	vs		3			g		CM	■8	o	GO	V\	VS		Os	Os		vs
	*3		uojsuaj qsfj	-3 CM	ns CM	CM CM	r-	8	vs CM	o CM	CM	8			■3	CM	CM CM	VS CM	vs CM	vs CM	CO	ns CM	sO CM	JZ	sO	8	co
	o														o												
			(emo/33Q spcy uo uoysuej,	§	8	?	8	8	£	&	s	8			;	s	s	8	8 CM	8	§	§	§	8	g	§	§
	*3		ojqBjsoJp^] 1SN	sO	ns		ns	CM		-3	CM	7			rd	vs	vs	VS	•3		-3	ns		T“			CM
	o														O												
	o		(ew/3>i)												O									s			
	a		aUJjpeN BIO J J	8	8	8	8	8	8	8	g	g			S3	8	8	g	g	g	8	8	g	§	g	§	g
	§	Sujsjay epcy uo		co	-3	Os					so	o	8 ©		g	co	-3	CM	O	vs	VS	K	co	cn	co	Os	o
	o	uojsuaj, ojne^sojpXH		SO	o	CM	ns	Os	-3	-3	Os	CM			■3 o	vs	n-	Os	•-		-3	o	-3				SO
	Os												i		Os												
	I	■p 8	poq.oejao0	8	8	8	8	§	_8	8	8	8	g		ft g 7j	8	8	£		§		£	£	§	£	§	I
	rd			c-	8		-3	co		US	co		2			vs	sO	Os	o		-3	Os	CM	6	vs	co	s
	O	(O'—' ©				CM			CM			w	E?		o												***
	CO ft	ft © & £	peq.oeajooufl	8	R	8	8	8	R	8	R	R	w 2 ft		co i	8	8	8	8	8	8	R	8	R	8	R	
	1	£3		-3	sO CM	VS	.3		Os	ns	o	-3	ft			-3	CO	ns	r*~	oo	-3		o	6a			
	o			CM		ns		CM	ns		CM	CM	i		o o		s	CM	CM	ns		CM	ns	r-	CM	CM	o
	r-		(euio/3x)										w s		n-												1
	g		&inq.0Bj£ q.B	3	3	O	8	8	8	8	o	8	•p		S3	8	8	*o	8	8	8	3	8	g	&	s	8
			TOJy tbutStja	«-	ns	ns	CM	r*	-3	CM	s	CD	§			CM	SO	ns	I*-	o-		C3	CM	o	co		vs
	■3 o	uo sssa^g ©SejaAy											■g S3		o o												
	SO		(e”®/3*) pwj utiurpcpyi	3 CM	8	8 CM	§	g		8 co	8 r-	8 sO	•p &		sO	s	§		8 o	§	g sO	s -3	C> CM	8	e	8	& SA
			eeuy x^uf^WO	O	r-	CM	o	CM	«-	O	o	r-			rd	Os	os	r-	CM	o	Os	O	o	Os	o	»-	O
	o	uo sseaq.g aS-RiSAy											i		O												
	vs		(V/°V) “8ox	ns	Os	Os	GO		co	vs	r~	t—	M		VS	$ Os	C*“	sO	vs	Os	CM	r*-	ns	VS	co		Hi
			ajnqoeja qe	&	r- vs	C—	sO -3	3	ns cm	co sO	8	-3 Os	M M		s		-3	8	3	-3 Os	3	-3 co	-3 r-	CM	3	K	-3 O
			trpBjqs T««BvpeN	o		CM	6		CM	o	4	4	©		rd	O				CM	d	4	CM	o	4	el	ns
	o												i		O												
	-3		(quaoued)	co	ns	r-	-3	-3	-3	sO	Os	n-			-3	co	ns		CM	CO •	co	CM	sO	CM •	n-	VS	ns
		eejy jo uojjonpajj		5	ps	co co	r- ns	£	Os CO	Os -3	CO n-	vs co			*1	8	s	3	CM co	<3 Os	CM VS	-3 oo	ns Os	-3 CM	£	CM co	£
	o														o												
	ns		V/°T	CM		-3	8	«	c*-	CO	Os				ns	VS	vs	Os			CM	-3	-3	CM		sO	
		eeuy tbuj^ oq. e©jy		sO	co	co			ns	Os	r-~	o				vs	•	SO	SO	’’Z	•	ns	VS	ns			o
	i—5 O	pjuja-puo JO oj^bh		CM	-3	co	*-	-3	Os	*“	-3	c-—			*3	CM	ns	ns	vs	Os	CM	sO	VS		-3	VS	
	O														o												
	CM		(ew/330 sqsej, jo												CM												
	1			8	88 vs «-		n o J3	Ss § 5s S ns Os n- ns		V) 1						(0 o 1			8§	o	§88		§§	I§§§§			
	o		©Subh ©Jnseejj	£ <SJ	us r*~	co ns		vs so	co ns		VS so	co ns			*3		r- CM	cm ns	VS so	co ns	43	vs e*-	co ns	•p •<	SA sO	r* co	co ns
	o														o												
	o		uearpedg		M			H							1				H			H				N	
	3 o		JO UOJ^BJUSJUO												3 o												<
Table IV. Summary of tension tests under pressure on plate 17P-
	CS £ s 3 o	jpSN jo eajy teqoj, oq eanqoej^ exysuej, jo ueJ v jo ojq.ua		03 O O	.128	§	co I 2 I I 0				1 1			
	CJ													
	£ §		(buo/3j1)	O ft CM	£	8	%	8	0 0					
			amqoeJj qu		CJ	sO	ft	CS c-	Os CJ					
	o o		uojsuej qSM	CJ	CJ	CJ		CJ						
	n 11		(eW0/33j)	O		300	8							
	g	spcv UO U0TSU9J,		ft	§			co	§.					
	3	ojqeqsojji^H qaM		SO	CJ	*-		sO	7					
	o													
													•P	p
													£	a
	o		(gUD/S:>i)										0)	©
	£	Su-ppsfi wojj		0 ft	g	g	8	8	8				B	e
	§	Sutstjv stxv uo			Os	us	-Sr	<5	US				0)	©
		irnsuex ojquqsojp^H		sO				sO	-£				£	£
													£	£
													TO	TO
	Os											c*	nJ	nJ
	£	■gg		8	8	8		8	§			8	©	©
	g	tn't-	peq.o9ujoo	_p xo	-£ O\	_3 us	us	us 0	~1	c*	c-	us	B	
	o				T-	CJ		CJ	r-					
	°	£											£	£
	co	S £										0	O	O
	£	© p		Q	8	0	2	0	g			8		
	g	£ ~p A s	p9q.o9juooun		ft	s	0	0	0	c*	c*			
	o	£			CJ	CS	-	CJ	r-				£	£
	u												nJ	nJ
													rd	rd
	r-		(giao/Ssi)								c*		£	£
	£			ft	§.	ft	0	ft	ft		8	8	to	hO
		USJV XUUjSjJQ		CJ C~		US CJ	sO Os	os O	cs co	c~	0 0	us Os	©	©
	o	uo	SS9jq.g 93BJ9AV										£	£
													£	£
													•rl	•rd
	so		f.. 00/351)			ft	ft		ft			8		
	£ g	p^07 nwirp<Bjcr q.13		0 CJ	8			ft			8		O	O
		way T«uj2jjq		CO Os			CJ 0	c~ CJ	c~ 0	O	0	CO Os	O	O
	o o	UO	SS9*iq.g 93T2J9AV										EM	Eh
	us		(v/°v) ®Soq	03	OS	sO	us	sO	US	O	CJ	cs		
	1		sjaqoujy	8	G	us	CJ sO	co c-	0 c-	CJ co	8	~P CJ		
	rd		uyejqs jujnquN	r-	r-	CJ	O				0			
	O													
	-£		(qusojed)	us		US	sO	cs	SO	O		sO		
	■3	esjy jo uojqonpea		cs 0	0 co	CJ Os	ft	-£ us	ft	so us	'O	CJ		
	o													
	cs a		V/°V	-£■	US	us	c-	Os	CJ	c-	ft	C-		
	g	B9JV	juuja oq uejy	c-	O		co		0	CJ		CJ		
	rd	J'BUjSjJQ JO ojqua		CJ	US	cs	*“	CJ	CJ	CJ	T-			
	o													
	CJ										TO O			
	£		(8ud/3>i)	Atmos.	ft 8	8 8	0	TO O	8 8	8 8		TO O	ftft	Q ° ft ft
	§		sqsej, jo		-=J	co c-	■g	£	CJ CJ	Os M3	■g	£	CJ Os	
	■3		aSuuH sjnsseja		us r—	co cs	c	<JJ	US so	CO r-			US US	co 0
	o													
	£													
	g		ueuijoedg		W								tsJ	
	rd O o		jo uojququejJQ											
cs £ g rd O O	CS f*S d	c~ CJ	Os O	.06	CS O
CJ					
£			8	8	8
rd	Os	0		Os	c-
O					
1	g	8	8	8	g
7)	US	US	cs		*-
0					
0					
£ B	8	8		8	8
	us	c-	r-		US
0					
Os					
£ § rd	8	8	8	g	8
	CS	US	0	Os	us
O					
03					
£ s	g	8 CS	8 CJ	O 8	8
0	'O	Os	c-		-£
0				CM	
c—					
£ g rH	0	O		8	O -P
	US	US	cs	CJ	
O					
sO					
£ B	ft	8	ft	ft	ft
rd	Ox	Os	Os	0	0
O					
US					
	US	CJ	sO	ft	co
£ §	O	CJ	2		Os
7!	*-•	r-	CJ	CJ	CJ
0					
-£	O	Os	sO	r*-	CJ
'l	ft	c-	co co	Os	US Os
0					
cs	cs		co	-£	US
	Os	us	cs	Os	co
7)	CJ	cs	co	r-	Os
0					
CJ	TO O	§§	g 8	8 8	ft
rH		f- CJ	sO 03	03 «-	co cs
O					
£					
g			IX		
0 0					12
corresponding stresses at a natural strain of 3.0 are 27,300, 23,600, 27,700 and 23,700 kg/cm2. These latter figures are so similar that very similar ballistic behavior of the four plates is to be expected, other things being equal. Included in the "other things being equal" is homogeneity.
(d)	Increase of ductility with hydrostatic pressure. — We now turn from a consideration of the phenomena of plastic flow to the phenomena of fracture. In the second report it was shown that the ductility, as measured by the natural strain at fracture, increases linearly with hydrostatic pressure. In Figs. 3, 6, 9 and 12, the pressure at which the specimen was pulled is plotted against the natural strain at fracture. There is much scattering of the points, indicative of the lack of homogeneity of the material, but within the limits of error, the relation again seems to be linear. The first three plates — 18a, 13F, and 17F — are arranged in order of increasing inhomogeneity. That this is the proper order is shown by the way in which the points are scattered. For plate 18a the points representing the X-, Y- and Z-orientations all lie on the same line within the limits of error. For plate 13F the points of the X- and Y-orientations are roughly on the same line, whereas those of the Z-orientation tend to lie on another line corresponding to easier fracture in this direction. Finally, for 17F the points are much more irregular; only the points for the X-direction tend to lie on a line at all; the points for the Y-direction are badly scattered, and for the Z-direction the results are so capricious that measurements could be made at all only for two out of four specimens tested. It happened that these two measurable specimens were those tested under atmospheric pressure.
(e)	Principal stress components at fracture. — Columns 11 and 12 of Tables II to V determine the principal stress components at fracture. In fact, both rr and 99 have the value given in column 11 and zz has the value given in column 12. The three components could de plotted together in three dimensions to give- 13 -
a fracture surface. When the stress in the substance attains a
i
value corresponding to a point on the surface,, fracture occurs.
The strain at fracture is another parameter, varying from point to point of the surface. Special values of it are given in column h.
In generalj the stress components rr and 06 shown in column diminish with increasing hydrostatic pressure (column 2) or with increasing strain at fracture. One would expect a regular connection,, but if column 11 is plotted against column 2 or great irregularity will be found.- Fracture data are always capricious3 and the irregularity is enhanced here by the fact that column 11 is the small difference of nearly equal numbers. The greatest regularity is shown by the most homogeneous steel,, 6x1 . The fourth test on this plate is out of line with the others’ this may be plausibly explained by premature fracture due to a flaw. If this point is discarded; the others are fairly regular. If now rr (or ©6) for 6X1 is plotted against ZZ; a straight line is obtained.
This line may be extrapolated to the point where rr - zz; it occurs at about 10,,000 kg/cm2. The significance of the point of intersection is that here the total stress system which produces rupture reduces to a hydrostatic tension. This hydrostatic tension which produces rupture has often been called the cohesive strength; for plate 6x1 the cohesive strength thus appears to be about 10;000 kg/cm2.
We may obtain approximately the same result by another line of argument. The mean hydrostatic tension corresponding to any stress system is (rr + 69 + zz)/3. The mean hydrostatic tensions for the first; second; third and fifth tests on 6X1 are; respectively-; 10;h00; 10;200; 1Q;3OO and 10;£00 .kg/cm2. These are constant; within limits of error — an interesting result in itself; doubtless of significance. Now when fracture occurs under pure _hydrostatic tension with no other stress component; the
A
strain must vanish. Hence; if we plot the four tensions just calculated against the corresponding strains and extrapolate to*■ ' - m -
zero strain, the corresponding hydrostatic tension will he the cohesive strength. But since the lour tensions are constant and hence independent of strain, the extrapolated value when strain vanishes is.the same constant'value,. or approximately 10,000 kg/cm2, checking with the result just obtained by another method. This figure for the cohesive strength is lower than would be estimated from previous speculations in the literature . The lowering of the figure is a result of the correction for necking.
(f)	Character of the fracture. — Finally, we consider the character of the fracture itself.. In the second report it was shown that the cup and cone fracture which normally occurs under atmospheric pressure tends to disappear at high pressures, the -area of the tensile part of the fracture — that is, the flat bottom of the cup as distinguished from the sides.'where the fracture is on shear planes — occupying a progressively smaller fraction of the total area of the neck, and vanishing altogether at pressures between 13,000 and 20,000 kg/cm2. Tlie last columns of Tables IT to V and Figs. U, 7, 10 and 13 show the fraction of the total area occupied by the tensile break as a function of the hydrostatic pressure during pulling. The-result already found is suostantiatedg the ratio of .the areas drops off linearly with increasing pressure, vanishing in the neighborhood of 13,000 kg/cm2, a pressure above which the fracture is entirely along shear planes.
The greatest effects of difference of orientation and lack of homogeneity appear when one studies the character of the fracture.
The X-direction for all plates — that is, the direction parallel to the face and parallel to, the direction of rolling — permitted satisfactory measurements. The Y- and Z-directions for 16A also permitted such measurements, showing approximate homogeneity, whereas the Z-direction of plate 13F and both Y- and.Z-directions of plate 17F in general gave such indefinite fractures that they could not be analyzed into a tensile and a shearing part.
I
Plate 17F in particular gave highly abnormal fractures. Mention
may be made especially of the specimen of 17F in the Z-direction

<
GO
LU
H
<
-J
CL
CM
d»
c c c o o o
o « « • « « w h. t~
■o -5 T> X > M
O 4<J
30.000
M3»N jo D9JV 0| ajnoojj aiisuax jo osjv jo oijoy
<
co
LU
H
<
_l
Q_
O»
iT
u u Q
•o ’o >- M
04-<1
o30.000
guiy 8x‘ B-injaojj 10 •••■US •«•>! P»J3»JJ0Q
Ll.'
ro
ui
H
<
O o
X >-
o +
o»
iZ
»P8N (oojjv 01 ajnpojj 8|isu»i )0 oa-iv jo O||oy
l£
ro
LU
H
<
_J
Q_
<0
o»
iZ
« u
TJ T> > IM
+ <
{
aoodoc30,000
paN JO OOJV O| ®jnj3DJJ P o®JV |0 Of joy
M
Fig. 12. PLATE 6X I19 -
a
which was pulled at a pressure of 11,000
'cm
In.the course
<
Fig. 1h. Double neck on a tension specimen from plato 17F.
of the experiment it was thought from the nature of the results that the specimen had been pulled to fracture.; but, on opening the apparatus, separation was found to be incomplete, an axial section, through the neck having the appearance indicated in Fig. 1h —■ an open channel through the center with two isthmuses on either side . This configuration
would appear to be the result of a combination of high ductility in the sound parts of the metal, imparted by the pressure, and of.strongly segregated impurities, nonductile and probably-, nonmetallic, such as slag. The-location of the inclusions is such as to- show up most strongly for specimens with the Z-orientation, .'but the effect of their presence may be seen also in the fractures of the X-orientation.
The shearing-part of these fractures never showed the- clean-cut slip planes characteristic of sound metal, but the surfaces of shearing slip were dotted with minute pits, giving a matt appearance.
■ The pressure at which the' shearing fracture disappears is distinctly lower for the X-orientation

of plate 13F than for the other two.
In homogeneous plate this may be a significant feature, because one might expect that the total energy to produce fracture would be greater in a material in which the tensile — that is, the brittle — fracture disappears at low compressive stress. This would indicate that plate 13F should be better ballistically than- 20
the other two. However, other disturbing effects arising from lack of homogeneity would appear to be so great as to mask any effects of this kind.
In studying the effect of pressure in changing the character of the fracture a number of photographs have been made of longitudinal sections through the tensile specimens of two of .the plates of the second report. The section was made by machining away half the specimen. For convenience in machining and mounting for photographing} the second shoulder of the specimen was first machined off. Figure 15 shows one of the sectioned specimens mounted for photographing. Figures 165 17, 18 and 19 are enlarged views of the neighborhood of the fracture for four specimens from plate Clh which had been ruptured at pressures of 1 atm, 5,000^ 10^000 and 15,000 kg/cm2., respectively. These show in the first place the progressively greater reduction of area as the pressure at which the fracture occurs increases — that is, the increase of ductility with increasing pressure. In the second place} the change in the character of the fracture is clearly shown. The "tensile” part of the fracture is the approximately plane surface perpendicular to the axis; this progressively diminishes in extent, both absolutely and relatively, with increasing pressure, until at 15,600 kg/cm2 it has entirely disappeared and the fracture is all "shearing" in character} with a multiple cone of shear at the center. It is especially to be noticed that the tensile part of the break is situated at the smallest part of the neck,, whereas the shear runs into the outer surface where the neck is larger.
5. Conclusion
The final conclusion to be drawn from these measurements} as far as ballistic applications are concerned, is one which,, if it had been anticipated^ would probably have made unnecessary the method of attack of this paper. The present measurements disclose ample reason for the difference in ballistic behavior of the different plates,, but this difference does not depend on anythingFig, 16. Specimen from plate Cli» broken in tension under atmospheric pressure.
xl
;>ii

- Z3 -
Fig, 17. Specimen from plate C12| broken in tension under a hydrostatic pressure of U000 kg/cm2.
Fig. 18. Specimen from plate C1li broken in tension under a hydrostatic pressure of 10,100 kg/cm2.
Fig. 19• Specimen from plate Clli broken in tension under a hydrostatic pressure of 15,800 kg/cm2.J
<
'.,e 7 , ■ f'	’ . , ‘ ‘ ■■ t
- 25 -
as subtle as a difference in the effects of hydrostatic pressure on the physical properties. Rather, it depends on lack of homogeneity due to large scale inclusions} the effect of which can be adequately shown by tensile tests under ordinary conditions at atmospheric pressure on specimens cut in different orientations.
In other words3 the bad plates are simply to be described as made from ’’dirty11 steels. This seems to be the conclusion that is also being arrived at from other lines of evidence. It does not follow of necessity that all dirty steels will give plates of bad ballistic propertiesj but., on the other hand., if a plate Is bad ballistically it is pretty likely to be made of a dirty steely in the absence of such obvious characteristics as too brittle a temper3 and so forth.
These present experiments show that the effect of pressure in increasing ductility carries over to dirty steels. The effect is differential’ the ductility of the sound part of the metal is increased more than that of the inclusions so that the apparent inhomogeneity may be increased by hydrostatic pressure. A natural expectation would be that fractures in plates made of dirty steels would be initiated at points closer to the projectile than in plates of sounder steels.

d
*
*'