Design and Construction of Deep Underground Basing Facilities for Strategic Missiles: Report of a Workshop Conducted by the U.S. National Committee on Tunneling Technology, Commission on Engineering and Technical Systems, National Research Council. (1982)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Nuclear Weapons Effects and Results of Previous Tests JOSHUA L. MERRITT Merritt CASES, Inc. Red-lands, California SUMMARY: The damage done at a given distance from ground zero at Hiroshima and Naga- saki would be inflicted by modern weapons at distances perhaps ten times as great. Also, dramatic improvements in the accuracy of delivery systems for missiles make it probable that a near-surface target (such as an egress portal) would be within the crater produced by any weapons used to attack it. Furthermore, the numbers of poten- tial attacking weapons are such that using proliferation of targets (such as egress portals) as a means of protection is not economically feasible. Protection, however, can be afforded in a deep basing facility by burying the facility deep enough to provide a suitable distance between the burst point of the attacking weapons and the facility. The weapons effect of most concern is the stress induced in the rock, propagating to the deeply buried facility. Results of tests in rock, from which these stresses can be inferred, indicate that such an approach may be feasible. The structural damage observed in several completely contained nuclear events (namely, events "Hard Hat," "Pile Driver," Mighty Epic," and "Diablo Hawk") yields data useful for planning the design and construction of deep basing facilities. Al- though much of it was inflicted on sophisticated, super-hard structures at high stress levels, some unlined and rock-bolted structures survived impressive stresses in the rock. The rock types included granite and tuff with a wide range of uncon- fined strengths and angles of internal friction, as measured by tests of conventional cores. I have been asked to summarize our experience over the last thirty-five to forty years in weapons effects, and particularly weapons effects on deep underground structures. I must do so in an unclassified nature. The experiences I shall cover, or at least touch upon, are our experi- ences in Hiroshima and Nagasaki. The attack on those two cities involved so-called "nominal" bombs, a nominal bomb being 20 kilotons of explosive energy. (The Texas City ship explosion in l944, incidentally, was esti- mated to be the equivalent of two to four kilotons of explosive energy.) I will then go into, in an unclassified way, a discussion of the nuclear weapons effects, emphasizing cratering, stress with depth, and what we know about the stress with depth. Finally, I shall very briefly go over what we learned from a series of experiments entitled "Hard Hat" in l963, 27

28 "Pile Driver" in l966, and—more recently—"Dining Car," "Mighty Epic," and "Diablo Hawk" beginning in l975. I mentioned the Hiroshima and Nagasaki experiences. Figure l is from a book entitled Effects of Nuclear Weapons, the first edition of which was produced in l946; there have been several editions since then, the most recent being in the l970s. These photographs show what happened at 0.5 mile from ground zero at Nagasaki (Figure 5.34a) and what happened 0.3 mile from ground zero at Hiroshima (Figure 5.34b); there was total destruction at those points. I mentioned this is a 20-kiloton nominal bomb. The yields of the bombs we are talking about today are in the neighborhood of 20 megatons, a thousandfold as great. To a reasonable degree of approximation, what happened at Hiroshima and Nagasaki at about 0.5 mile would occur at 5 miles from our current weapons. It is an awe- some amount of energy and an awesome amount of damage that can be created by that energy. Colonel Berry has already mentioned CEP (circular error probable). Dr. Sevin has mentioned stress with depth. Figure 2 is a cartoon which I borrowed from Air Force Systems Command Manual 500-8, published in l967. I have added some rough outlines to emphasize some of the points that we need to at least touch on. The most important point is the cra- ter created by a surface or near-surface burst of a nuclear weapon. If that burst should occur at or near the surface of a very competent rock —granite or basalt, as an example—the radius of that crater* is about 500 feet for a l-megaton device. If you take that up to current opera- tional sizes, we could multiply that by a factor of three. So, instead of 500 feet in radius, we are talking some l,500 feet in radius, about 0.25 mile for the radius of the crater. The depth of the crater, again, for l megaton for scaling purposes is something on the order of l00 to l20 feet. You scale that up to, let us say, a 27-megaton device, it be- comes 300 to 360 feet in depth. The accuracy of the weapon is such that if an enemy aims at a target, he can almost certainly place that target within the crater. For soils, to jump to another extreme while not at- tempting to imply any solution in terms of siting, the crater, instead of being some l,500 feet, could be on the order of 3,000 feet in radius. I marked also on the figure "EMP" and "prompt radiation." I will not go into any depth on those. Suffice it to say that EMP (electromag- netic pulse) is the most awesome lightning strike that you could imagine multiplied by many, many-fold. The prompt radiation is also a signifi- cant item and could create significant damage to anything on the surface. As the stress waves propagate downward from the crater, we have the directly induced ground shock, which Dr. Sevin has already touched upon. Figure 3 shows our experience in hard rock on the left. The first four are granitic sites. The French data is in granite for weapon yields of 3.6 kilotons to ll7 kilotons. Hard Hat, in l963, was conducted in granite at the Nevada Test Site (Climax Stock granite) with a yield of 5.9 kilotons; "Shoal," north in Nevada, was again at a granite site with *Here we are referring to the apparent crater, that which exists after fallback has occurred. The true crater and the associated rupture zone may be much larger.

29 a yield of l2.5 kilotons; Pile Driver, 59 to 6l kilotons in l966; and the last three on this chart are for andesite at Amchitka, Alaska, rang- ing from 8l kilotons for the "Longshot" event up to a 5-megaton device for the "Canikin" event. Plotted within the two bars is the summary of all the measurements of particle velocity in those particular shots and then, using an acoustic impedance to relate particle velocity in feet per second to stress in kilobars, we have a separate bar on the ordinate for the stress in kilobars. The 0.5 kilobar used by Dr. Sevin would corre- spond to a scaled distance below a contained event on the order of 700 feet for a total confined explosion. The preponderance of experiments that we have conducted in the United States have been in the hard rock and the tuff at the Nevada Test Site. Most of the tuff is at Area l2, Nevada Test Site. On the right-hand panel of Figure 3 we show the scatter bands of data from the left-hand panel. Superimposed on the right-hand panel is the measured particle velocity from experiments in tuff. You can see that the lower bound of the data for hard rock becomes essentially the mean for the data in the softer rock, specifically tuff. Again using an acoustic impedance to convert particle velocity to stress, you find a lower stress in tuff as compared to granite. Dr. Sevin has already men- tioned the coupling. I would emphasize that the data shown in the figure are strictly from contained bursts. We have to convert from contained bursts to the conditions of a surface burst by use of the coupling factor mentioned by Dr. Sevin. We made up Figure 4 in cartoon form to summarize our data base for behavior of lined and unlined openings in rock in the United States. The underground explosion test series conducted in l948 to l953, logistically supported out of Dugway Proving Ground, Utah, included granite, limestone, and sandstone, with a tunnel below the burst point. The burst point was a buried burst; much of the data were gathered by documenting the behav- ior of those tunnels following the detonation. They were all chemical explosives, ranging in size from 320 to 320,000 pounds. The sizes of the tunnels went from 6 feet in nominal size for a modified horseshoe up to 30 feet in size. The 30-foot tunnel was subjected to the effects of a 320,000-pound burst. The 6-foot tunnels were subjected to the effects of a 2,560-pound charge or, in a few instances, a 320-pound charge. I have flagged the test sites that I have already mentioned and the sizes of weapons: 8 pounds to l60 tons for the UET (underground explosive test) series; the nuclear events go from 55 tons to 5 megatons, 5 mega- tons being for the Amchitka shot. The series of experiments have in- volved Hard Hat and Pile Driver, as already mentioned. First is a car- toon of these events which I will discuss in greater detail a bit later. Next is a cartoon of the Mighty Epic/Diablo Hawk events that were con- ducted in the middle to late l970s. Finally, we summarize the peak stress of up to a kilobar (a kilobar is 14,500 psi) for unlined cases. For lined cases we have experienced all the way up to 5 kilobars (or 72,500 psi) stress in the rock. Final- ly, in the table we summarize the types of linings, the environments, and the materials in which we have conducted experiments. The basalt, mentioned at the bottom of the chart, incidentally, was at the Nevada

30 Test Site; the salt is located in two places. We have done free-field experiments in those media also. Let us move to the Hard Hat event (Figure 5). The Hard Hat event was reached through a shaft 785 feet deep to the muck pocket; the muck pocket went an additional 35 feet below the intersection with the nearly horizontal drift. There were 3 experimental stations and some 43 test structures in these 3 areas. The device was emplaced in a 36-inch cased drill hole some 943 feet below the surface; as already indicated, it was a 5.9-kiloton device. The working point, as we call it, or the zero point, was depressed below the structure's drifts to get rid of the shad- owing that might occur from one drift to another if they happened to be at the same elevation. The plan view of those three drifts, A, B, andC, is shown in Figure 6. "A" drift was some 250 feet from the zero point. "B" drift was 340 feet and "C" 460 feet from the zero point. The 5.9-kiloton device was to the left, off the figure in this sketch. There were l0 structures in A drift, l8 in B drift, and l5 in C drift. The basic design was for the conditions estimated to occur in B drift, and then the structures were arrayed at three different locations in order to give a spectrum of dam- age. Stress levels inferred from measured particle velocities at A drift were 2 to 4 kilobars, and at C drift, 0.5 to l kilobar. A series of mainly cylindrical structures were involved, ranging from the strongest structure, a reinforced concrete structure 8 inches thick surrounded by 20 inches of polyurethane foam, to the weakest of the structures, a horseshoe shape with 4-inch, l3 pound-per-foot steel shapes with 2-inch lagging between the shapes. I will not have time to go into any great detail on Hard Hat, but I think from the slides I shall show on Pile Driver, subsequently, we can infer some of the conditions that occurred in the Hard Hat experiment. Now, moving to the Pile Driver experiment (Figure 7), I shall show a perspective with the access shaft some l,367 feet deep, extending to a muck pocket 89 feet deep, and then some l,400 feet along the access drift to a winze. The winze goes down some l04 feet; the device was placed at the bottom of it. The device was planned to have a 50-kiloton yield. It actually turned out to be a 59-kiloton yield and in some references it has been noted as a 6l-kiloton yield. The test structures were located in X drift, at 320 feet from the zero point, on out to C drift at some 940 feet from the working point. Measured particle veloc- ity at X drift was sufficiently high that it corresponds to about 30 kilobars—about 500,000 psi—in the rock, on out to about l0,000-20,000 psi, or 0.66-l.33 kilobars, at the most remote range. From the perspective, you should note that we varied the size of excavation from 44 to 7 feet in size. We also varied configuration: X intersections, T intersections and complete structures, capsules at the bottoms of the X intersections. The structural types included rock bolts, unlined openings, and various types of sophisticated lining, but before we touch briefly on the construction methods and the results of that particular experiment, I would like to note some of the major fea- tures of the geology at Area l5 of the Nevada Test Site.

3l In Figure 8, the plan is just reversed over the preceding perspec- tive. The lines are the surface maps of the various major joints encoun- tered throughout the workings. The joints were mapped at the tunnel lev- el, some l,400 feet below the surface. At that level, the contact between a quartz monzonite and a granodiorite was as shown. The physical proper- ties of the rock types were almost identical, but one was a much more quickly cooled material than the other. Also at tunnel level, we had a horsetail fault that we picked up a definite expression of at the base of the shaft and near C drift, but we did not pick it up clearly in B drift. The granite was a jointed rock and it did have some faulting and discontinuities in it. Some damage along natural joints can be seen in Figure 9. The dark- er areas in the roof in the foreground represent regions where small blocks of rock fell. Figure l0 is a post-test picture of a more sophisticated structure. This one is seven feet in internal diameter. It is 6 inches thick, has nominally 0.5 percent reinforcement on each face in the circumferential direction and 0.25 percent reinforcement on each face in the longitudi- nal direction. The "flex duct" used to provide air was installed after the reentry; during the event itself the opening was completely free of materials. The power line was also brought in for electric power after the event. The only things that existed within this structure at the time of the event were the signal cables, which were strapped to the wall with airplane cable in one case and with bungee cord in the other case. You can see the bungee cord in place. Surrounding this seven-foot struc- ture was some four feet of material, frequently referred to as Merlcrete. It is a foamed neat cement that has a flat-top stress-strain curve. That structure survived somewhere between 0.66 and l.33 kilobars. Other struc- tures actually survived at a level of two to four kilobars, as I shall show in this next slide. Figure ll shows a steel structure, but there is a concrete structure very similar to the one we just saw in the background in this particular view. This figure is in B drift. B drift saw a measured particle veloc- ity of about ll0 feet per second, which, depending on how you want to convert that into stress, is somewhere between 2 and 4 kilobars. The concrete structure in the background survived. The steel structure in the foreground used corrugated steel of two thicknesses. It was surround- ed by four feet of the foamed neat cement. It also survived two to four kilobars. Again, there was a power line and a "flex duct" that were put in after the event to give us ventilation and power. On the left rib of the structure are the cables for getting the instrumentation sig- nals out. They were held down with bungee cord or with airplance cable. They were covered with spray-in-place foam to further protect them. I mentioned a rock bolted section. Figure l2 shows a heavily rock bolted section. The rock bolts are some two feet on centers. There are at least two layers of chain link fence on the surface. The rock bolts are size number ll; they are l6 feet long. The opening is l6 feet in diameter. This picture was taken after the event, and there is no evi- dence of any distress whatever in that particular structure. I would hasten to add several things, however. First, this is an end-on

32 configuration. (The stress wave propagated in the direction of the lon- gitudinal axis.) The working point, if you eye down the rock bolt with the white painted bearing plate, is some 846 feet ahead of that. The best estimate of the stress level at that particular point is about l.5 kilobars, about 20,000 psi, in the rock. Also, I would emphasize that these rock bolts are very closely spaced. These rock bolts were also tensioned to 60,000 pounds force in each of the bolts after they were in place, so that there was a fairly high confining stress intentionally put on the rock. Finally, in my last five minutes I shall try to bring us up-to-date with the recent series of tests and, because I am running out of time, let me try to expedite this by first quickly indicating in plan view the Mighty Epic event (Figure l3). The Mighty Epic event was originally planned as a line-of-site (LOS) pipe experiment, as they are called, with test chambers off the view on the right side to test other effects of the device. Of course, that device creates a stress wave that propagates out- ward and we took advantage of that stress wave and added a series of structures in the Mighty Epic event. Mighty Epic had as its main thrust looking at so-called super-hard construction. The Diablo Hawk event following it reloaded that super- hard construction in a second loading. In passing, let me briefly com- ment on the result of a reloading of a structure where it first saw one kilobar propagating in a direction perpendicular to the longitudinal axis (side-on) in Mighty Epic; and axially in Diablo Hawk with stress levels, depending where you were in the drift, anywhere from one kilo- bar to about three-eighths kilobar range. There was some distortion of the interior steel ring resulting from the second loading, but the actual measured distortion was on the order of one-half inch. The Mighty Epic working point appears on the left. In Diablo Hawk, we not only reloaded the structures that you saw in the previous case, but we added a number of other experiments. One was a sand-filled tun- nel to determine the behavior of potential underground reservoirs and finally, there was a series of size-effects experiments at 0.6, 0.3, and 0.l5 kilobar. The size of the last set of structures ranged from 9 inches to l8 feet. Construction of the horseshoe-shaped drifts for size effects was with a roadheader in a single pass for the 9-foot and l3-foot openings. Two passes were required for the l8-foot excavation. The completed structure was some 54 feet long. The openings were largely unlined. There were rock bolts between the various unlined segments, so that in the event that one of them failed, it did not propagate to the next seg- ment. There was some dislocation of the rock from the back in several places in the unlined openings, but probably nothing to cause any great concern about moving personnel and equipment through that tunnel sub- sequently at the lowest stress level. In closing, I would like to emphasize that high-speed photography, taken in those types of drifts, which you will see later today, shows what appears to be an awfully hostile environment, but there was no se- rious damage in many drifts. Assuming, of course, that one has provided secondary protection for anything that might have been housed in the

33 drifts, I would not be concerned about surviving the environment shown in the high-speed photograph. So, my time is doubtless up; I shall close at this point. SPEAKER: Jay, what is the stress level for that unlined tunnel? DR. MERRITT: The unlined tunnel shown in the last slide saw 2,400 psi. We had two other drifts of similar size that saw 4,000 psi and 8,000 psi. The 8,000 psi one was very heavily damaged. The one at 4,000 psi was moderately damaged. This one was lightly damaged. SPEAKER: On a scale of one to ten, at the present time, what is our re- liability on the test related to instrumentation cabling, etc.? Where would we stand in analyzing where we go from here on that reliability? DR. MERRITT: Gosh, I am not quite sure how to answer that because, as you are probably aware, we are actively considering the development of new gauges to allow us to go into still higher stress levels than we have looked at there. There is a lot of work in being able to do that. The instrumentation for most of the things I have shown here, up to one kilobar in tuff, has survived quite well, with some exceptions. There is some faulting in this particular rock, and we did have one case, and I emphasize one case, in which a fault did guillotine our instrumenta- tion system. SPEAKER: In our time span of collectively analyzing and going into a basing mode, do you think that between now and l984 there is enough re- search and development being done in this field that should bring us to a better percentage of reliability? What I am referring to is one of our discussions has to be what is coming out of new technology, and is it sufficient? DR. LINGER: I think, in answer to that question, the instrumentation technology has improved probably two orders of magnitude since these tests were conducted, and I think that we will see a high-speed photo- graph of this tunnel during the shot, and the man that is going to do that is the manager of the Construction Division at Nevada Test Site, and I think he has tremendous reliability built into his instrumentation, more so now than ever before. SPEAKER: What was that rock, the last one we were looking at, the un- lined tunnel? DR. MERRITT: The unlined tunnel there is a tuff at N tunnel, approxi- mately l,400 feet below the local surface. The unconfined compressive strength of that particular rock, using NX cores, is in the neighborhood

34 of 3,000 to 4,000 psi. If you want to characterize it as a Mohr-Coulomb material, it probably has an angle of internal friction in the range of l0 to l5 degrees. Its specific gravity is right around two. Its seismic velocity is about 8,000 feet per second. INDUSTRIAL STRUCTURES 167 :v • FIGURE 1 (5.34a) Destroyed industrial area showing smoke- stacks still standing at 0.5l mile from ground zero at Nagasaki (from The Effects of Nuclear Weapons, l977).

35 168 STRUCTURAL DAMAGE FROM AIR BLAST ; •• FIGURE 1 (5.34b) A circular, 60 feet high, reinforced- concrete stack at 0.34 mile from ground zero at Hiroshima (from The Effects of Nuclear Weapons, l977).

36 v; 1 V) Ul u 0 *» u 3 +J +J 4J ^ & •o 0) •H U 0 U) 0) 4J U 01 o 0 •a o M TD 10 -U CO <d M •H M-l 0 CM LIJ a:

37 la £ 5 *3 c Q I* o ui CD Q- U. UJ (C UJ w O u. o t>ll 11 I I—I tllll I I 1 I I.II 9 / ll»ll II I It 1 1" " • • / •- •* 2 > in'111 i—i—ii..... ? > s i i n« I t i -25 3mm *- ^£

38 <u on 03 CQ 03 -4—' £ c cs ^c; CT>-P r-1 — 4) C C/0 0. 'C-uO O ui 1 a. u. UJ OC UJ M D u. Z Q Z O tc. O II O oc ^1 X 4 UJ i O O OB -.13 2?> s ° 2 U uJ «! z o' E cd IA UJ k H X U O I O </i — _j a. M UJ O 11! o - S « < « t; « < " i£ ffl H ^ * • I o — UJ UJ U 5i UJ UJ UJ c ac = O UL UJ Z O _! UJ > u o cc ^ 2 2 ="£ o S'l si I I id Is •^x i el M + \\ ** UJ Q£

39 0> SURFACE lS» \i 36" CASED HOLE (STATiOM UlSoI 800' ELEVATOR SHAPT (STATION iSOOl TEST T -A ELEV OF -t TEST DRIFT-A TEST DRIFT -B TEST 1 -C ." IH ZERO POINT - Vertical Section. Event Hard Hat FIGURE 5

40 — TEST DRIFT c - Plan View of Test Sections. Event Hard Hat FIGURE 6

4l CO to u o: 3D CD

42 CO UJ cc CJ

43 FIGURE 9 Failure along natural joints, Shop Drift (event Pile Driver). FIGURE 10 Structures with packing, sections CRl and CRla (event Pile Driver).

44 FIGURE 11 Structure with packing, section BR9 (event Pile Driver). FIGURE 12 Structure with rock bolts and mesh—end-on loading, sections CR5 and CR6 (event Pile Driver).

45 CO LU OL