D. Siskind, V. Stachura, M. Stagg & J. Kopp
RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining STRUCTURE RESPONSE AND DAMAGE PRODUCED BY AIRBLAST FROM SURFACE MINING David E. Siskind, Geophysicist Virgil J. Stachura, Geophysicist Mark S. Stagg, Civil engineer and John W. Kopp Mining engineer Twin Cities Research Center, Bureau of Mines, Cities, Minn. ABSTRACT The Bureau of Mines studied airblast from surface mining to assess its damage and annoyance potential, and to determine safe levels and appropriate measurement techniques. Research results obtained from direct measurements of airblast-produced structure responses, damage, and analysis of instrument characteristics were combined with studies of sonic booms and human response to transient overpressures. Safe levels of airblast were found to be 134 dBL (0.1 Hz), 133 dBL (2 Hz), 129 dBL (6 Hz), and 105 dB C-slow. These four airblast levels and measurement methods are equivalent in terms of structure response, and any one could be used as a safe-level criterion. Of the four methods, only the 0.1-Hz high-pass linear method accurately measures the total airblast energy present; however, the other three were found to adequately quantify the structure response and also represent techniques that are readily available to industry. Where a single airblast measuring system must be used, the 2-Hz linear peak response is the best overall compromise. The human response and annoyance problem from airblast is probably caused primarily by wall rattling and the resulting secondary noises. Although these will not entirely be precluded by the recommended levels, they are low enough to preclude damage to residential structures and any possible human injury over the long term. INTRODUCTION Airblast, like ground vibrations, is an undesirable side effect of the use of explosives to fragment rock for mining, quarrying, and excavation. Blasts at large surface mines and quarries can produce noticeable airblasts at large distances, particularly when weather conditions are favorable for propagation. Because of these variations in propagation, and the strong relationship between blast confinement and airblast character and levels, prediction and control are often more difficult for airblast than for such other adverse blast effects as ground vibrations, dust, and fumes. Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 1 1 of 85 This report summarizes research by the Bureau of Mines on airblast effects on residential structures. Discussed is research by the Bureau and other institutions on ground vibration response and damage, human response, sonic booms, airblast generation and propagation, and instrumentation as they apply directly to the airblast-tolerance problem. Reports are being prepared on blast-vibration generation and propagation, ground vibration damage, and instrumentation methodology, and while work is continuing on many other aspects of the blasting problem including blast design and human annoyance. Research in areas related to airblast was also analyzed-specifically, sonic booms and human response to transient overpressures. Most of this work is in general agreement with the Bureau's results; however, it was mainly supportive data because of characteristic differences in the sources and their resulting effects. An understanding of how residential structures respond to airblast and the airblast characteristics most closely related to this response will enable blasts to be designed to minimize these adverse effects. The mining industry needs not only appropriate design levels for blast effects, but also practical techniques to attain these levels. At the same time, environmental agencies responsible for blasting control and noise abatement must be provided with reasonable, appropriate, and technologically established and supportable criteria on which to base their regulations. Finally, neighbors around mines and other blasting operations require protection of their health and property (fig. 1). ACKNOWLEDGMENTS The authors wish to acknowledge the very generous assistance of regulatory agencies, engineering consultants, powder companies, homeowners, and mine and quarry operators. Paul D. Schomer of the Corps of Engineers and George W. Kamperman Associates made helpful suggestions in the area of measurement techniques. Special thanks are due to the Illinois Environmental Protection Agency for demonstrating the immediate need for this airblast research. Much of the fieldwork and data reduction was done by student employees Alvin J. Engler, Steven J. Sampson, David M. Purdham, Patrick G. Corser, Stephen Lepp, and Michael P. Sethna, who put in long hours. Philip D. Murray, whose Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 2 2 of 85 thesis material was referenced in the structure-character analysis, also provided additional support. AIRBLAST CHARACTERISTICS Causes of Airblast Airblast is an impulsive sound generated by an explosive blast and resulting rock fragmentation and movement. Four causes of airblast overpressures are generally recognized: (1) direct rock displacement at the face or mounding at the blasthole collar, (2) vibrating ground, (3) gas escaping from the detonating explosive through the fractured rock, and (4) gas escaping from the blown-out stemming. Wiss labels these four contributions to the total airblast (1) air pressure pulse (APP), (2) rock pressure pulse (RPP), (3) gas release pulse (GRP), and (4) stemming release pulse (SRP) (83)4. Their characteristics have been described in various other studies (53, 58, 83). The GRP is also termed the gas vent pulse (58). The air pressure pulse (APP) will dominate in a properly designed blast, and will only be absent for cases of total confinement (that is, underground blasts). Each blasthole acts as an APP source. Close-in or front-of-face airblast measurements with wide-band systems usually detect a series of APP pulses corresponding in time to the interval between the top decks or front-row holes. At large distances or behind the face, dispersion and refraction mask the individual pulses and the blast timing becomes less evident. The time histories then lose their APP spikes and associated high frequencies. The rock pressure pulse (RPP) is theoretically generated by the vertical components of the ground vibration summed over all the area, which acts as a large vibrating piston. A simple relationship was found by Wiss (53, 83) between RPP and the vertical ground vibration Vv: RPP = 0.0015 Vv, with RPP in pounds per square inch (lb/in²) and Vv in inches per second (in/ sec). Normally, RPP has the least amplitude of the airblast components; however, it is typically of higher frequency (identical to the Vv which spawns it), and enables us to predict the minimum airblast level expected (for example, 1.0 in/sec Vv will generate 0.0015 lb/in² , or 114 dB-peak). It arrives at the receiver simultaneously with the ground vibration and prior to APP. The gas release pulse (GRP) and stemming release pulse (SRP) are the most undesirable and theoretically controllable parts of the airblast, since they involve the blast design variables of stemming, spacing, burden, and detonation velocity. SRP and/or GRP result from a blowout and appear as a spike or series of spikes superimposed on the APP. Because they have rise times of only a few milliseconds, they are rich in unwanted high-frequency airblast energy. Snell (58) reports that simply the use of an AN-FO explosive contributes to the irregular occurrence of SRP because of its slow detonation. Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 3 3 of 85 Other conditions that may contribute to this effect are small-diameter holes (lower detonation velocities), wet holes, long columns, and high propagation velocities of the rock. Consequently, SRP would be more of a potential problem for quarries than coal strip mines. Figure 2 shows a coal mine Production blast soon after initiation. The mounding which produces APP energy and the stemming plume are both visible, signifying that less than total confinement was obtained. Surface detonating cord is a potential source of high-frequency airblast, and at small to moderate distances may be the dominant source. It is easily controlled by increasing the ground cover, and its effects diminish with distance. Airblast Types Observed In Mining Airblasts from surface mines have been classified according to their frequency character (53). Figure 3 shows the time history and spectra of a type 1 airblast which has prominent APP pulses resulting from almost line-ofsight propagation conditions, and exhibits a 15-Hz spectral peak corresponding to the 60-msec separation between hole detonations. This 15-Hz peak in the spectra is not the largest, but it is the most important in terms of its noticeability and effects on structures. The magnitude of the APP peaks is a fundamental result of the rock fragmentation process, and cannot be appreciably reduced. However, the delay interval and the resulting airblast frequency are part of the blast design and can be controlled. A type 2 airblast is shown in figure 4, with the APP pulses spread out into a single, very-low-frequency overpressure. This type of airblast typically occurs at large distances and behind the rock face. For quarries, APP pulses are produced by rock movement directly away from, and in front of, the face. The relatively high frequency airblast energy represented by the APP spikes cannot readily diffract behind and around obstacles, including the face itself. Consequently, type 1 airblasts are typically encountered in front of the face, and type 2, behind. An exception to this noted by Stachura (61) involved a high face across the pit from the blast. The face served as a simultaneous reflector and high-pass filter and returned the APP pulses as a ghost type 1 airblast. For coal mine highwall shots in area strip mines, where little or no rock displacement occurs, the heaving of the bench at the collar of each hole generates some Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 4 4 of 85 APP, which should not be as horizontally directional as it is in contour mines or quarries. For all blasts, the air is a dispersive and selectively absorptive medium for sound transmission. The high frequencies are attenuated at a higher rate, and all airblasts become similar to type 2 at large distances. The time history and spectra of a coal mine highwall shot producing a blowout and significant SRP appear in figure 5. This sharp pulse caused a large structural response and a high level of sound. Theoretically, blasts can be designed to prevent the generation of SRP and GRP; however, the natural variability of the blasted material (mainly, its nonhomogeneity and anisotropic character) makes it impossible to control SRP at all times. Small blasts such as those used in construction and coal-mine-parting shots are particularly troublesome, not only for the high levels of airblast they can produce, but also because they are of high frequency (as much as 5-25 Hz compared with the usual 0.5-1.5 Hz). Obtaining sufficient confinement is the usual problem with these shots. Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 5 5 of 85 Unconfined Blasts Even more serious than poorly confined blasts is the problem of totally unconfined blasts exemplified by artillery, open-air detonations, uncovered surface detonating cord, and explosive testing. These produce high-frequency airblast and the highest levels per amount of explosive. Studies of the effects of unconfined airblast cannot readily be applied to the mining airblast problem, except possibly to provide a worst case or, when unconfined blasts are observed at large distances, to simulate confined blasts (58). These studies are discussed in the "Human Tolerance" section. Sonic Booms A typical sonic boom time history (N-wave) and spectra are shown in figure 6 (86). Considerable work has been done on the damage from and response of structures and humans to sonic booms. With caution, these results can be applied to the blasting problem. The period of a sonic boom depends on the aircraft size and ranges from 75 msec for an F-104 to 206 msec for an XB-70. The spectrum is smoother than an airblast and like it contains much low-frequency energy. Sonic booms do not have isolated frequency spikes as do SRP and APP, and probably should not be directly equated in effect to type 1 or blowout-dominated airblasts. Most sonic boom spectra drop off at 12 dB per octave in Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 6 6 of 85 pressure from the spectral peak, which can be roughly determined by inverting the N-wave duration and typically ranges between 4 and 11 Hz. MEASUREMENT AND INSTRUMENTATION Airblast is a transient time-varying overpressure, which can be expressed in any units of pressure. Various types of studies have specified pounds per square foot, pounds per square inch, millibars, and Newtons per square meter, and various expressions of relative sound levels, in decibels (dB). An equivalence and conversion chart for overpressure units is shown in figure 7. Sound Pressure Levels Shown in figure 7 is a line representing the sound pressure level (Lp) defined by the standardized relationship: Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 7 7 of 85 Survey Instrumentation The measurement and recording systems used for the Bureau of Mines airblast studies have been described in interim reports (54, 55). Low-frequency pressure transducers of 0.1- to 380-Hz response were used in 7- and 14 channel FM recording systems (figs. 8-9). From these "ultralinear" airblast time histories, other "linear" measurements were generated by appropriate filtering. The 0.l-Hz low-frequency response was required for research purposes to measure accurately the l-Hz energy often present in the airblasts (8, 53, 56). The high-frequency response of the measuring system could be a problem for some sources (detonating cord, SRP), although in practice, only a 200-Hz response is required (23). The 0.l-Hz airblast time histories were processed by playback through various analysis systems (including the filtering networks of standard sound-level meters,) and then correlated with measured structure responses. Supplementing these values were direct measurements using a 0.18,000-Hz sonic boom measuring system (B&K 2631) and sound level meters giving Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 8 8 of 85 2-Hz, 5-Hz, 6-Hz linear, and C-weighted-slow values. The analyses are further described in the section on processing airblast time histories, and also in Stachura's report (61). Structure responses and ground motions were measured by direct-reading velocity gages of 2.5- and 4.75-Hz natural frequencies (Vibra-Metrics 120 and 124) with flat frequency responses of 3-500 Hz and 5-2,000 Hz (-3 dB), respectively (62). The airblast measuring instruments and their application (table 1) are discussed in other reports (5, 38, 54, 61). It is often convenient to measure airblast with blasting seismographs, most of which have an airblast channel as well as three components of ground vibration. They typically give permanent film or paper records, but often limit the choices of weighting, integrating times, and frequency ranges. Stagg (62) and Stachura (61) describe these systems, many of which have been frequency-calibrated by the Bureau of Mines. Two of the devices in table 1 are not complete systems, but transducers which require some type of recorder (B&K 2631 and Validyne DP-7). Two are impulse precision sound level meters with multi-function capability (B&K 2209 and GenRad 1933). Permanent records can be obtained by using a suitable recorder on their outputs; however, the sound level meters give only numerical readings. The B&K 2209 has a "hold" capability which greatly facilitates the reading of transients. The acoustic monitor (Dallas AR-2) is designed for long-term unattended recording. The ultralinear system is the only one which accurately measures the true waveform, and should be used wherever later processing is required. Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 9 9 of 85 TEST STRUCTURES A total of 56 different structures were studied for airblast and ground vibration response and damage (table 2). All were houses, except No. 54, which was a mobile home. In addition, structures 13, 15, 16, and 50 were somewhat larger than single-family residences. Some structures (19 and 20) were studied for a variety of blasts, highwalls, parting, and surface. The response of structures 1-6 were described in an earlier study (55). Of the 56 structures, only 17 had significant and identifiable levels of airblast response (figs. 10-24). In many cases, the blasting did not result in high airblast levels and/or high-frequency airblasts. Measurements were generally made near the blasts since ground vibration were also being sought. Time separation between the ground vibration and the airblast was not always sufficient to identify the latter response. The coal-mine-parting and quarry shots usually produced good airblast data, as did the coal highwall shots with long delay intervals. Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 10 10 of 85 Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 11 11 of 85 FIGURE 10. - Test structure 12, metal mine. FIGURE 11. - Test structure 14, metal mine. Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 12 12 of 85 FIGURE 14. - Test structure 21, coal mine. Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 13 13 of 85 FIGURE 15. - Test structure 22, stone quarry. Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 14 14 of 85 Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 15 15 of 85 Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 16 16 of 85 Instrumenting For Response Outside ground vibration, airblast, and corner and midwall responses of the structure were measured for each shot. The ground vibration was measured by three orthogonal 2.5-Hz velocity gages buried about 12 inches into the soil next to the foundation (62). Outside airblast was measured with at least one DP-7 gage, and two sound level meters (one reading C-slow). The structures were instrumented for horizontal motions by a pair of Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 17 17 of 85 gages mounted low on the first floor vertical walls in the corner closest to the blast and one or more midwalls. Typically, the vertical motion was measured in the same corner. Additional channels were usually available and used for various additional corner motion measurements at mid-heights, near the ceiling, or on the next floor; additional floor-motion measurements such as mid-floor verticals; basement wall horizontal measurements; opposite-corner responses (for rotational motions); and inside noise. Corner measurements assessed the racking motions (distortion) of the structure. Essentially all blast damage occurs where stresses and deformations are produced within the planes of the wall as shear stresses. Consequently, the vibration measurements made in the corners were assumed to indicate damage potential, because they measured whole-structure response. Other types of response caused different but consequential results. Midwall motions (perpendicular to the wall surface) are primarily responsible for window sashes rattling, picture frames tilting, dishes jiggling, and knickknacks falling. Midwall accelerations in excess of 0.4 g (12.8 ft/sec~) are occasionally generated and could cause items to fall off shelves. These midwall motions are not necessarily dangerous to the structure since walls can vibrate in this mode without producing high levels of stress. Midwall motions are mostly annoying. Floor motions present a problem similar to midwalls. Like them, they also produce secondary noises and can lift hanging objects off nails and cause them to drop to the floor. Structures are designed to resist normal vertical load, so vertical corner motions of less than 1 g should not warrant serious concern. Copyright © 2000 International Society of Explosives Engineers RI 8485 Structure Response and Damage Produced by Airblast From Surface Mining - P 18 18 of 85 Natural Frequencies and Damping Natural frequency and damping are the most important structure-response characteristics. The natural frequencies of the structures as measured from blast-produced corner motions are summarized in figure 25, with individual values listed in table 2. Structures continue to vibrate after the sources (ground vibration and airblast) decay, and natural frequencies and damping can be measured from the time histories. The vibrations of structures, especially midwalls, are approximately sinusoidal; therefore, the natural frequencies are calculated by inverting their periods (in seconds). The damping values are given by B = 100 2Πm Ln (An /An+m) , where B is the percentage of critical damping, A is the peak amplitude at the n~th cycle, and m is any number of cycles later. Murray (28) discussed the general problem of structure frequencies and damping and also computed many of the values in table 2. He noticed that damping values were level-dependent, indicating that friction was nonlinear.