Canal Excavation Blast Safety Research Proposal

Page 1

ia iv it Sandia National Laboratories Albuquerque, New Mexico 87185-0175 U.S. Department of Energy Oakland Operationi Office Site Management Division, L-57 Lawrence Livermore National Laboratory Subject: Your Request for Review for Releasibility of Sandia OUO Report March 1, 1963, subject: Canal Excavation Blast Safety Research Proposal, Document Marked OUO (Official Use Only) Per your request Bruce Green and I have reviewed the referenced OUO Sandia document, provided by you, to determine if it contains any information that is still considered to be sensitive. Our review determined that the document contains no sensitive information and is publicly releasable. We have removed the OUO markings to assist you, and the document is enclosed. Enc: Referenced Document Exceptional Service in the National Interest

Page 2

Chlamember 30, 1995, this document was reviewed by Richard B. Cramer, SICL Classification Officer, !lr'RATORY Dept. 7447, and was determined to be publicly LIVERMOiZE Nevada Operations Office Las Vegas, Nevada ORIGINAL SIGNED BY Original Si fr,ed by Re: Canal Excavation Blast Safety Research Proposal The enclosed proposal describes research for calendar years 1964, 1965, and 1966, associated with feasibility studies of operational safety from blast effects of nuclear excavation for a sea level canal. Anticipated costs are included. Two previous proposals have been submitted, one for related research during 1963, submitted December 31, 1962, and one for a site survey during 1964 and 1965, submitted January 28, 1963. This completes the three-phase proposal set as you have requested.

Page 3

CANAL EXCAVATION BLAST SAFETY RESEARCH PROPOSAL To describe the research program associated with the operational safety from blast effects of nuclear excavation for a sea level canal along Routes 17 ,and 25. Period to be covered is from January 1, 1964 until Proposed excavation along Route 17 would require about 300 devices, with a total yield of 170 mt. The maximum yield for one device would be 15 mt. From 35 mt to 50 mt may be fired at one time. Route 25 excavation would require about 650 devices, totaling 330 int, the largest device being 10 mt. As much as 50 mt may be fired at one time. Sedan shot demonstrated that there is an off-site blast problem associated with large yield cratering shots. Two microbarograph stations in California recorded pressure amplitudes only ten percent smaller than would be expected "at the same range and yield air burst in an homogeneous atmosphere. Scaled to 50 mt, this propagation could break windows out to 280 miles range. There are four cities of 100,000 population, or more, within 280 miles of Route 17 and seven within the same range of Route 25. About five million people live within these radii. Atmospheric sound ducting can cause blast concentration at long range. Blast predictions are currently subject to considerable errors although refinements are continually being developed. The magnitude of the canal excavation effort and its dependence, in some respects, on blast safety, makes further refinement necessary and economically justifiable. There are four categories of uncertainty in assuring blast safety. Chronologically, in order of occurrence from a shot, they are: 1.) source function description, 2.) propagation mechanics, 3.) prediction techniques, and 4.) blast damages. Current models with their unknowns and uncertainties, and suggestions for improvement in each area are set forth in what follows:

Page 4

Canal Excavation Blast Safety Research Proposal -2- March 1, 1963 1.) Source Function Description: First, we define air blast transmissivity as the ratio of the air blast pressure amplitudes recorded from underground or cratering shot to the air blast expected at the same range from the same yield burst in free air in a homogeneous atmosphere. Data from cratering experiments have shown apparent increases in transmissivity with distance from a shot. An air blast source model developed to explain this consists of a strong blast wave initially confined within a conical section above burst. Sides of the cone may conform approximately to the true crater cone. Along the ground outside but near the crater only relatively weak blast pressures are recorded from energy diffracted down from the strong shock sector. At increasing range, where enough time for diffractive effects has elapsed, transmissivity is observed to increase gradually. Successively higher elevated shock rays within the cone will travel higher into . the atmosphere and upon refraction back to ground.level land at larger ranges and thus demonstrate increasing transmissivity with range. Relative shock strengths along rays elevated 20 to 30 degrees above the horizon are apparently conserved after carrying through the ozonosphere and on return to ground near 100 to 150 miles range. To date, individual station measurements at this distant range have shown transmissivity factors which vary from about 4 to 40 percent for optimum cratering burst depths. Results from Plowshare events Stagecoach, Buckboard, and Scooter, and from Danny Boy and Sedan show several variations which need explanation. Transmissivity appears to increase with yield, it appears to increase with ground density (from comparison of hard rock and desert alluvium burst environments), and finally, it appears to be reduced for nuclear versus high explosive sources. Measurements of transmissivity to date have shown considerable scatter mostly because of meteorological fluctuations. Pressure signals separated by only a few seconds in time of arrival at one station may show transmissivities which appear to vary by factors of two or more. This uncertainty makes relation to yield and burst environment questionable. To date, blast prediction concern with blast sources has almost completely neglected row charge effects. Some recordings were made

Page 5

Canal Excavation Blast Safety Research Proposal -3- March 1, 1963 on Toboggan tests but this data collection has not been analyzed as yet. These data need to be digested and augmented with measurements from larger yield row charge tests to allow source model development for ditch digging. Close-in blast measurements in horizontal and vertical planes through test crater bursts are needed to verify physical reality of our assumed source model. Measurements .on test shots of various yields and in various soils would help in extrapolating to source functions for yields and soils involved in canal excavation. Acoustic propagation with some modification has been generally assumed. That is, sound ray path calculations for refraction in an atmosphere of horizontally stratified wind and temperature gradients determine where various sound ducts will give ground level sound waves. Calculated ray densities or spacings then provide 'a measure of relative blast perturbation amplitudes. The free air burst reference standard for close-in blast parameters (above .37 psi) in a homogeneous atmosphere is the IBM Problem M. Empirical curves for weapon tests differ from IBM Problem M because of non-uniform atmospheric conditions which surrounded actual bursts. Beyond the .37 psi contour, it is assumed that the overpressure distance relation is Ap-1.2 This exponent was derived from experiments with negligible atmospheric refraction. Losses apparent in this law (as compared to acoustic decay conditions) exceed losses which might be expected from finite amplitude shock wave propagation theory. If these losses were caused by classical molecular viscous absorption, they would be frequency and thus yield dependent. Experimental data have indicated independence of yield. Sound wave interactions with atmospheric turbulence may be the real attenuating mechanism but this effect is not well understood at present. Further theoretical study and probably additional experiments are needed to adequately explain this empirical propagation law and assure its yield independence. A redundancy of observations of propagations in various tests, not only nuclear cratering shots, is needed to assess first the accuracy of mean blast estimates and then the statistical probabilities of various

Page 6

Canal Excavation Blast deviations from this mean. This statistical variance from short-term wind turbulence is needed for estimating prediction error distributions. Gravity wave oscillations of the entire atmosphere are set off by large yield shots (megaton range). These waves may have safety implications for very large cratering bursts and the pressure-distanceyield relations for this propagation mode should be determined. Theoretical solutions for gravity wave propagation, such as the iteration-variation method of Byatt and DeVault, may be usable for prediction but computation details need to be worked out. Yield, depth of burst, and burst environment are used to scale the IBM Problem M reference curve back and forth and up and down on a pressure-distance graph for homogeneous atmosphere propagation. Variations around this curve are used for azimuth angles and elevation angles as described in the non-uniform early wave model. Atmospheric ray path calculations then proceed. Calculations are based on Snell's law for refraction of rays passing through media with sound velocity gradients. The input is a set of temperature and wind data at various significant altitudes in the atmosphere. Long-range propagation calculations for large excavations will generally require inputs to altitudes of nearly 200,000 feet. Ray calculations are prepared for selected azimuths of interest. An analog sound ray computer for this purpose is operated at NTS CP-1 for operational predictions. The same calculation has been programmed for high-speed digital computers available at Sandia Laboratory for post-shot analysis Output sound ray patterns may take several weather dependent a.) If, in a particular direction, sound velocities at all altitudes are less than at the ground surface, all rays will be turned away from earth, through the ozonosphere, and pass on into the ionosphere. Only diffracted wave energy will return to the ground; it will cause no

Page 7

Canal Excavation Blast Safety Research Proposal -5- March 1, 1963 blast damage except close-in to construction and site facilities. b.) Sound velocity may increase with height in a shallow surface layer to give an inversion sound duct. Waves carried by this duct are rapidly attenuated by surface roughness and diffuse reflections. These waves also are of interest only for on-site and relatively close-in predictions. These predictions for construction facilities, however, should be refined by better measurements of reflection coefficients at low incident angles. Too, general inversion propagation problems could be studied using the iteration-variation wave solution methods. c.) Sound velocity at some height well above the surface may be higher than at the surface and intermediate levels. In this case, a portion of the early blast wave will be turned back into a ground level sound arc or ring some distance from the burst. In temperate latitudes, jet stream winds may duct sound into a caustic region at 30 to 100 miles range. These jet winds are rather rare in equatorial regions. In the ozonosphere, however, temperatures are nearly as warm as at ground level so winds may cause ducting to east or west, i.e., downwind of the ozonosphere wind direction. Ozonosphere winds generally blow from the east in summer, and from the west in winter. Day-to-day variations in these winds cause considerable variability in blast pressures recorded at any point in the sound ring, which ordinarily ranges between 60 and 150 miles. This duct is a most important consideration for establishing off-site blast safety from the proposed canal excavations. Nevada test data show that, at 135 miles range downwind of ozonosphere winds, average pressure amplitudes from 20-kiloton shots are 2 millibars and scatter with a factor of 2.5 probable error. Scaled to 50 megatons, this would mean nearly 50 millibar average pressure amplitudes which would cause considerable damage. A factor of 10 attenuation (or 10 percent transmissivity) from a cratering shot, still allows damaging pressures about two-thirds of the time. At double this distance, the probability of damaging shocks would be nearer 10 percent. To be assured of safety, upper winds should be measured and ray calculations made operationally integral to the shot or firing decision.

Page 8

Canal Excavation Blast Safety Research Proposal -.6- March 1, 1963 Calculated concentration or focusing of rays striking the ground allows, first, blast pressure predictions by their relation to pressures from the spherically divergent source treated by the standard pressure curve. Transmissivity factors are then applied to make actual and more believable pressure predictions. Practice on Nevada underground tests and high explosive calibration shots has given encouraging results. There are frequent gross errors. Statistics for these errors should be collected so that future predictions can have specified accuracy distriAnother reason for scatter and error is that blast waves may be quite sharply focused by refraction. Acoustic calculations show sharp foci exist where a wave front folds over to give infinite energy concentrations. Such concentration cannot be supported in air, of course, since pressure waves have finite thickness. Very strong local amplifications have, however, been observed at Nevada in sound patterns ducted by jet winds. Hydrodynamic shock calculations should be attempted to resolve amplifications that are physically possible in air. Increased statistical collections should be made of Nevada data by increasing the number of high-explosive calibration shots associated with other tests. Since atmospheric variability effects are so large, fairly numerous collections are necessary to ascertain the Nevada experience has shown that large windows begin to break when blast pressure amplitudes exceed 4 millibars, on typical Nevada-Utah towns. Observations of cracking plaster have been made when pressure amplitudes were 3 millibars, although most claims for these damages were unjustified. There have been many occasions when towns in Nevada have been subjected to more than 2 millibars pressure amplitudes with little complaint or response. Tropical cities, with fewer large glass windows, may be less sensitive to blast than our own U. S. cities. This should be established by inspection and consultation.. Also, some response and acceleration measurements on real housing structures should

Page 9

Canal Excavation Blast Safety Research Proposal -7- March 1, 1963 be made along with microbarograph measurements of Nevada tests in hopes of relating structural plaster damage response to weak blast pressures. Comparative readings should be made for stresses from natural phenomena such as winds and commonly withstood influences such as heavy trucks passing and.doors slamming. From these, comparative values could be derived in laymens language for public relations. It is expected that a system for observing and forecasting meteorological parameters, calculating blast propagation, applying yield and burial attenuation factors, etc., will be in hand to allow blast amplitude predictions which will verify somewhere within * 20 miles of a prediction point. Statistics will be available so that yield, depth-of- 'burst, and season of the year may be selected with regard for the surrounding population centers and structures for a prescribed shot delay probability. Further statistics should be prepared so that a specified probability and level of damage will not be exceeded by blasting a particular crater or ditch section when the current atmospheric conditions are approximately known from field measurement. a. Description of Program: Things to be done in the 1964, 1965, and 1966 period are conveniently broken into categories of 1.) theoretical research, 2.) collection and analysis of simulation data from field research experiments and device tests at NTS or other testing sites, and 3.) on-site survey of pertinent conditions in the Panama region. The first category will require only man-hours-scientific expenditure and, thus, will be limited principally by available talent. The second will require funding to some extent with total costs beimg apportioned between Plowshare research funds, test site safety requirement funds, and Panama requirements funds. Third category work was covered separately by the Phase II proposal for on-site surveys.