# A Designer's Approach for Estimation of Nuclear-Air-Blast-Induced Ground Motion.

1. IntroductionUnderground siting of strategic structures [1] is an option to enhance the safety against nuclear-air-blast [2]. Nuclear explosions generate rapidly moving air-overpressures capable of producing significant ground displacements [3]. Most severe loading occurs within a close vicinity around the ground zero (GZ) known as superseismic zone. This zone is subjected to significantly high overpressures. The velocity of moving air-over-pressure fronts is also more than the P-wave velocity of the ground [3-6]. Hence, a reliable estimate of nuclear-air-blast-induced ground displacement is required for design in superseismic zone [5, 7, 8]. Several studies using numerical approaches that account for realistic stress-strain behaviour and boundary conditions are reported for calculating the free-field response [9-16]. Such studies require specialized expertise in numerical tools and are unattractive to practicing engineers. Before the advent of computational tools, Wilson and Sibley [17] conceptualized a one-dimensional pseudostatic approach to estimate air-blast-induced ground displacement.

Whitman [4] and Baron et al. [18] have also shown that air-blast-induced ground motion is predominantly one-dimensional and vertical in superseismic zone. In the present work, a generalized one-dimensional pseudostatic formulation is developed to estimate nuclear-air-blast-induced vertical ground displacement that accounts for (i) nonlinear stress-strain behaviour [e.g., 19,20], (ii) stress-dependent wave propagation velocity, and (iii) stress wave attenuation. Using the proposed formulation, a closed-form solution is developed for linearly decaying blast load with negligible rise-time applied on a layered ground medium with bilinear hysteretic stress-strain behaviour. The proposed model is validated against the field data of an atmospheric nuclear test conducted at Nevada test site [21], and various parametric studies are carried out. In addition, performance of the closed-form solution is also compared with the model given in UFC [22].

2. Generalized Formulation

The developed formulation is explained in detail below.

2.1. Determination of Overstress Distribution (due to AirBlast) in the Ground. Air-overpressure generated due to nuclear-air-blast consists of an initial rising portion followed by a decaying portion (Figure 1). Overpressure time-history can be visualized as consisting of multiple overpressure fronts. These overpressure fronts are transferred to the ground as P-waves. On arrival of an air-shock front at the point of interest, an initial pulse travels with seismic P-wave velocity and reaches a depth [Z.sub.0i] at time [t.sub.i]. Subsequently, other fronts arrive and penetrate into the ground. If the location of peak overpressure front at time [t.sub.i] is given as [Z.sub.pi], then, the ground above the depth [Z.sub.pi] experiences compressive stresses lower than those caused by the peak overpressure front at that depth. Therefore, the depths above [Z.sub.pi] are referred to as "unloading zone." However, the peak overpressure front is yet to reach at locations deeper than [Z.sub.pi], and therefore, the zone lying beneath the depth [Z.sub.pi] is denoted as "loading zone" (Figure 1). It is noted that "unloading zone" corresponds to the "decay portion" of overpressure, whereas "loading zone" corresponds to the "rising portion." Therefore, at times earlier than the rise-time, there does not exist any unloading zone as overpressure fronts from the decay portion do not arrive before the end of rise-time.

To determine the stress distribution in the ground at time [t.sub.i], contribution from all stress fronts arrived before time [t.sub.i] is considered. If the kth overpressure front arrives at the point of interest on ground surface at time [t.sub.k] such that [t.sub.k] [less than or equal to] [t.sub.i], then the kth overpressure front reaches at depth [Z.sub.ki] at time [t.sub.i]. Stress fronts propagating through ground are affected by attenuation (caused by hysteresis losses, viscosity of ground materials, and dispersion of energy in 3-dimensional space) and interference with reflected wave fronts (generated due to impedance variation with depth). In the present study, stress wave attenuation is taken into account through a geometrical attenuation parameter, whereas interference of incident and reflected stress waves is ignored. If the attenuation coefficient at depth [Z.sub.ki] is given as [alpha] ([Z.sub.ki]), then the stress generated at depth [Z.sub.ki] can be written as a([Z.sub.ki])x P([t.sub.k]), where P([t.sub.k]) is the magnitude of the kth overpressure front. Depth of the penetration [Z.sub.ki] depends upon wave propagation velocity of the stress front, which in turn depends upon the tangential modulus of the geomaterial at the stress level at the depth of interest. If wave velocity of the kth stress front is denoted as [V.sub.kz] (a function of depth "z"), then (1) can be solved to obtain [Z.sub.ki] as a function of [t.sub.i], [t.sub.k], and [V.sub.kz] (2),

[mathematical expression not reproducible], (1)

[Z.sub.ki] = g([t.sub.i] - [t.sub.k], [V.sub.kz]) = [phi] ([t.sub.k]), (2)

where g is the function obtained by solving (1). It is noted that for a given [t.sub.i], [Z.sub.ki] is effectively a function ([phi]) of [t.sub.k] only because [V.sub.kz] is also a function of [t.sub.k]. As soon as the air-blast slaps the ground surface, ground deforms elastically and an elastic wave is generated which propagates with seismic P-wave velocity. Upon subsequent arrival of higher overpressure fronts, inelastic waves propagate at velocities smaller than P-wave velocity. If the ratio of P-wave velocity to wave velocity corresponding to the kth overpressure front at depth "z" is denoted as [f.sub.kz] = ([V.sub.pz]/[V.sub.kz]), then at initial stress levels [f.sub.kz] is equal to one. With increasing stress level, [V.sub.kz] decreases [17, 23]; hence, it is logical to assume that [f.sub.kz] increases with increasing stress level and the maximum value is attained at a peak stress level corresponding to the peak overpressure front. Therefore, a simplest choice is to assume a linearly increasing [f.sub.kz] in direct proportion to overpressure (3). Overpressure fronts from the decay portion propagate through the media which is already stressed in a nonlinear range due to passage of the peak overpressure front. Therefore, it is assumed that wave velocity of overpressure fronts in the decay portion is the same as the peak overpressure front, and hence [f.sub.kz] is the same as [f.sub.rz]:

[mathematical expression not reproducible], (3)

where subscript "z" represents the depth coordinate. Using the average loading rate as [P.sub.o]/[t.sub.r], P([t.sub.k])/[P.sub.o] can be approximated as [t.sub.k]/[t.sub.r] (3). As wave propagation velocity through ground media is proportional to the square root of the tangential modulus [24], [f.sub.rz] is given by the following equation:

[mathematical expression not reproducible], (4)

where [mathematical expression not reproducible] is the tangential modulus at initial stress level at depth "z" and [mathematical expression not reproducible] is the tangential modulus at peak stress level at depth "z."

It should be noted that the ground is in equilibrium under geostatic stresses and at rest before nuclear explosion and only the overstresses caused by nuclear-air-blast causes the ground displacement.

2.2. Determination of Strain Distribution in Ground. Strain distribution with depth ([[epsilon].sub.iz]) can be determined using stress distribution ([[sigma].sub.iz]) and an appropriate stress-strain relationship [e.g., 25-27] as given in the following equation:

[mathematical expression not reproducible], (5)

Caption: Figure 1: Schematic representation of dilatational stress distribution in ground at time [t.sub.i]

where functions [f.sub.L] and [f.sub.U] denote the loading and unloading branches of the stress-strain curve, respectively (Figure 2).

2.3. Integration of Strains. To obtain the vertical ground displacement [u.sub.i] at time instant [t.sub.i], the strain distribution is integrated (6) from the ground surface to the penetrated wavelength in the ground up to time [t.sub.i]([Z.sub.0i]),

[mathematical expression not reproducible]. (6)

Substituting Z = [phi]([t.sub.k]) in (6) leads to the following equation:

[mathematical expression not reproducible]. (7)

Thus, the displacement time-history can be estimated using (7).

3. Closed-Form Solution

Using (7), closed-form solutions can be obtained for several simplified cases [28]. In this article, a closed-form approximation is developed for the following simplifications:

(a) Linearly decaying overpressure time-history with zero rise-time (Figure 3, (8)),

P([t.sub.k]) = [P.sub.o](1 - [t.sub.k]/[t.sub.eq]), (8)

where [P.sub.o] and [t.sub.eq] are peak overpressure and equivalent positive phase duration, respectively.

The reason behind choosing the special case with zero rise-time is the popularity of approximating the actual decay of the incidental pressure by an equivalent triangular pressure pulse among practicing engineers [8, 22]. It is to mention that design charts and empirical relations between the weapon yield, peak overpressure, and the equivalent positive phase durations are also available for equivalent triangular pulses with zero rise-time [5, 8, 29]. Thus, for practicing engineers, a linear decay model with zero rise-time is more useful.

(b) Bilinear hysteretic stress-strain model [8] with the loading secant modulus ([M.sub.L]) and strain recovery ratio (r) as parameters (Figure 2).

(c) Attenuation coefficient ([alpha]) is given by the following equation [8, 30]:

[alpha] = [(1 + z/[L.sub.w]).sup.-1], (9)

where [L.sub.w] - 0.971 x [W.sup.1/3] x [P.sup.-1/2.sub.o] x [V.sub.L],

where W is the yield of the explosion in kiloton and [V.sub.L] is the wave propagation velocity of the peak overpressure front.

(d) Using (4), [f.sub.rz] is determined as 1 for the bilinear stress-strain model. However, due to nonlinear stress strain behaviour, [f.sub.rz] is usually greater than 1. Wilson and Sibley [17] and Batdorf [23] recommended a range of 1.5 to 2. Hereafter, [f.sub.rz] is denoted as f and assumed to be constant with depth.

(e) A constant but representative P-wave velocity [V.sub.p] of ground media is assumed.

Using assumptions (a)-(e), integral for loading fronts in (7) (with [t.sub.r] [right arrow] 0) between general time instants [t.sub.x] and [t.sub.x+1] is written as [mathematical expression not reproducible], where L([t.sub.x] [right arrow] [t.sub.x+1]) can be shown to be given by the following equation:

[mathematical expression not reproducible]. (10)

It is to clarify that the overpressure time-history defined by the piecewise function P ([t.sub.k]) such that P ([t.sub.k]) = [P.sub.o][t.sub.k]/[t.sub.r] for [t.sub.k] [less than or equal to] [t.sub.r] and P([t.sub.k])= [P.sub.o] (1 - ([t.sub.k] - [t.sub.r]))/ ([t.sub.eq] - [t.sub.r]) for [t.sub.r] [less than or equal to] [t.sub.k] [less than or equal to] [t.sub.eq] converges to (8) in the limiting case when [t.sub.r] [right arrow] 0. Therefore, the expression L ([t.sub.x] [right arrow] [t.sub.x+1]) in (10) is first obtained for a case of finite [t.sub.r] and then it is evaluated for the case when [t.sub.r] [right arrow] 0 to determine the solution for overpressure time-history of (8). Similarly, the integral for unloading fronts in (7) (with [t.sub.r] [right arrow] 0) between general time instants [t.sub.x] and [t.sub.x+1] can be written as [mathematical expression not reproducible] where U ([t.sub.x] [right arrow] [t.sub.x+1]) can be shown to be given by the following equation:

[mathematical expression not reproducible]. (11)

By setting the appropriate values of [t.sub.x] and [t.sub.x+1] in (10) and (11), the closed-form integrals between desired time instants can easily be evaluated. Thus, the closed-form solution is given by the following equation:

[mathematical expression not reproducible]. (12)

The closed-form solution is further extended to accommodate multiple ground layers. This requires the determination of time instants [t.sub.x] and [t.sub.x+1] during which the wave fronts pass through a particular layer. For illustration, a layered ground medium with an interface at depth H with the modulus of top layer as [M.sub.1] and modulus of bottom layer as [M.sub.2] is considered as shown in Figure 3. This leads to the three cases (Table 1) and corresponding displacement solutions (shown in Table 1).

4. Validation

The proposed closed-form solution is validated against the field data from an atmospheric nuclear test conducted at Frenchman Flat (Nevada). Perret [21] presented the measured overpressure time-histories (Figure 4) along with corresponding peak ground displacements at different distances from GZ (Table 2) for a 37 kt nuclear explosion at a height of 214 m. Model input parameters for above mentioned nuclear test are determined as follows.

The recorded overpressure time-histories are converted to equivalent linearly decaying overpressure time-histories (Figure 4). Equivalent overpressure time-histories are given by (8) such that the peak overpressure ([P.sub.o]) is equal to the recorded peak overpressure, and equivalent positive phase duration [t.sub.eq] is given by the following equation:

[t.sub.eq] = 2[I.sub.p]/[P.sub.o], (13)

where [I.sub.P] is the positive phase impulse (area under the recorded overpressure time-history). It is worth mentioning, though the rise-time is taken to be zero in the equivalent overpressure time-history in (8), that the effect of non-zero rising time has been accounted for by considering the total positive phase impulse (in (13)) which includes the area under the rising and decaying portions of actual overpressure time-history. Therefore, it is expected that setting the rise-time as zero in (8) would have a negligible impact on the magnitude of peak displacement. Furthermore, the ratios of length of rise-time to total positive phase duration for the four cases (P1, P2, P3, and P4) are 0.10, 0.21, 0.23, and 0.27, respectively. This indicates that rise-time is reasonably small near ground zero; however, the relative length of rise-time increases with increasing distance from ground zero. Thus, it is expected that the errors (if any) due to setting rise-time to zero would be more pronounced only at distances far away from ground zero. Even in such cases, a closed-form solution can be developed using (7) by setting a finite rise-time. However, it is worth noting that at far away distances (from ground zero), other important factors related to outrunning ground motion would start governing [31].

The initial pulse moving at seismic P-wave velocity ([V.sub.p]) penetrates to a depth of [V.sub.p][t.sub.eq] for equivalent case, whereas for actual case, it penetrates to a depth of [V.sub.p][t.sub.p]. It is noted that the depth of the stressed zone in the equivalent case is smaller compared to the actual case as equivalent duration ([t.sub.eq]) is smaller than the actual positive phase duration ([t.sub.p]) (Figure 4). Thus, all depth-dependent parameters are scaled by a factor SF = [t.sub.p]/[t.sub.eq], such that the total depth of the stressed zone becomes SF x [V.sub.p][t.sub.eq] = [V.sub.p][t.sub.p] with the depth of the top layer being H/SF, and with a characteristic attenuation length of [L.sub.w]/SF. Based on the variation of P-wave velocity with depth at Frenchman Flat (Figure 5), the average P-wave velocity ([V.sub.p]) is estimated as 658.69 m/s. Whitman [4] presented the representative constrained modulus for Frenchman Flat interpreted from different experimental techniques (Table 3). Whitman [4] also emphasized that all aspects of stress-strain behaviour of geomaterials under blast loading are not captured by a single test and advocated to choose the modulus judiciously. The stress attenuation coefficient is calculated using (9) with scaled [L.sub.w] (Table 4). Using the closed-form solution (shown in Table 1), parametric variations are studied (Table 5) and peak displacements (Table 6) are estimated.

The parametric studies (Table 5) highlight that the proposed model is most sensitive to the constrained modulus compared to other parameters and choice of the modulus is also subjective for engineers. However, based on the parametric studies, the last column "Recommended Value" of Table 5 provides some guidelines to select the constrained modulus. With increasing availability of similar case studies, these guidelines can further be refined.

5. Comparison with UFC Model

The UFC manual [22] provides an expression for air-blast-induced peak vertical displacement based on one-dimensional elastic wave propagation:

[u.sub.i|max] = [I.sub.p]/[rho][V.sub.p], (14)

where [rho] is the bulk density of geomaterials. To estimate peak displacements using (14), positive phase impulse ([I.sub.p]) is taken from Table 2. Representative P-wave velocity ([V.sub.p]) and density ([rho]) are taken as 658.69 m/s and 1331 kg/[m.sup.3], respectively, taking into account the variation with depth (Figure 5). Computed peak displacements using the proposed model and UFC model are shown in Figure 9 along with the measured field values. It can be clearly seen that the predicted values of the proposed model are in good agreement with the measured values and the UFC model significantly underestimates the peak displacements under high overpressures (with increasing war head capacity).

6. Conclusions

A closed-form expression is developed to estimate nuclear-air-blast-induced free-field ground displacement that takes into account peak overpressure, positive phase impulse, depth of layer interface, representative constrained modulus of each layer, strain recovery, stress attenuation, P-wave velocity, and velocity ratio. The solution is validated against a nuclear test conducted at Frenchman Flat (Nevada) and the following conclusions are arrived at:

(i) Peak ground displacement estimates are quite sensitive to the constrained modulus, and a judicious selection of appropriate constrained modulus based on the magnitude of applied overpressure is recommended.

(ii) Based on the presented case study, two guidelines are recommended to select the appropriate modulus: (a) shallow ground layers are likely to have modulus values determined by unconfined or triaxial compression test and deep ground layers are likely to have modulus values computed from seismic velocity test, and (b) constrained modulus increases with decreasing ratio of applied stress to overburden.

(iii) Effect of attenuation should be accounted for low overpressures and may be neglected at higher overpressures (or high war head capacities).

(iv) The velocity ratio affects mainly the time of occurrence of peak ground displacement and not the magnitude under higher overpressures. Therefore, the velocity ratio becomes important when design calculations utilize the complete displacement time-history such as in case of shock spectra.

(v) A complete strain recovery is a better representation of actual conditions under low overpressure zones, and under higher overpressures, a partial strain recovery is recommended.

(vi) The proposed model closely estimates the experimental values at all overpressures, whereas the UFC model significantly underestimates the peak ground displacements at high overpressures.

Notations [alpha]([Z.sub.ki]): Attenuation coefficient at depth [Zk.sub.i] [[epsilon].sub.iz]: Strain at time ti at depth "z" [[epsilon].sub.pz]: Peak strain level in geomaterial at depth "z" [[epsilon].sub.rz]: Residual strain level in geomaterial at depth "z" f: Representative velocity ratio (between P-wave velocity and peak stress velocity) of ground media [f.sub.kz]: Ratio of P-wave velocity to wave velocity corresponding to the kth overpressure front at depth "z" [f.sub.L]: Functional form for the loading branch of the stress-strain curve [f.sub.rz]: Ratio of P-wave velocity to wave velocity corresponding to the peak overpressure front at depth "z" [f.sub.U]: Functional form for the unloading branch of the stress-strain curve H: Depth of the top ground layer [I.sub.p] Positive phase impulse of overpressure time-history [L.sub.w] Characteristic attenuation length [M.sub.L]: Loading secant modulus [M.sub.2]: Modulus of bottom layer [M.sub.1]: Modulus of top layer [P.sub.o]: Peak overpressure P([t.sub.k]): Pressure magnitude of the kth overpressure front r: Strain recovery ratio [rho]: Bulk-density of geomaterials SF: Scale factor [[sigma].sub.pz] Peak stress level in geomaterial at depth "z" [t.sub.eq] Equivalent positive phase duration of linearly decaying overpressure time-history [t.sub.k] Time when the kth overpressure front arrives at the point of interest on ground surface [t.sub.kH]: Time of arrival of the last loading front to reach at depth H (by the time tf) [t.sub.p] Positive phase duration of overpressure time-history [t.sub.r]: Rise-time to peak overpressure in overpressure time-history [u.sub.i]: Vertical ground displacement at time [t.sub.i] [V.sub.p]: Representative P-wave velocity of ground media [V.sub.L]: Representative wave propagation velocity of the peak overpressure front in ground [V.sub.kz]: Wave velocity of the kth stress front at depth "z" W: Yield of the explosion [Z.sub.ki] Depth penetrated by the kth overpressure front at time [t.sub.i] [[delta].sub.iz] Stress at time ti at depth "z" [Z.sub.0i]: Depth penetrated by initial pulse at time [t.sub.i] [Z.sub.pi]: Depth penetrated by the peak overpressure front at time [t.sub.i] ([partial derivative] Tangential modulus at initial stress [[sigma].sub.z]/ level at [partial derivative] depth "z" [[epsilon].sub.z])| .sub.[sigma]=0]: [mathematical not Tangential modulus at peak reproducible] stress level at depth "z."

https://doi.org/10.1155/2018/3029837

Conflicts of Interest

The authors declare that there are no conflicts of interest regarding the publication of this paper.

References

[1] S. Pinto, "Underground construction of nuclear power reactors," Nuclear Engineering and Design, vol. 61, no. 3, pp. 441-458, 1980.

[2] C. V. Chester and G. P. Zimmerman, "Civil defense shelters: a state-of-the-art assessment," Tunnelling and Underground Space Technology, vol. 2, no. 4, pp. 401-428, 1987.

[3] S. Glasstone and P. J. Dolan, The Effects of Nuclear Weapons, Castle House Publications, London, UK, 1980.

[4] R. V. Whitman, The Response of Soils to Dynamic Loadings; Report 26, Final Report, Massachusetts Institute of Technology, Cambridge, MA, USA, 1970.

[5] R. E. Crawford, C. J. Higgins, and E. H. Bultmann, The Air Force Manual for Design and Analysis of Hardened Structures, Final Report Jun 71-May 74 (No. AD-B-004152/5), Civil Nuclear Systems Corporation, Albuquerque, NM, USA, 1974.

[6] B. M. Das, Fundamentals of Soil-Dynamics, Chapter 4, Elsevier, Amsterdam, Netherlands, 1983.

[7] H. L. Brode, A Review of Nuclear Explosion Phenomena Pertinent to Protective Construction, R 425 PR, The Rand Corporation, Santa Monica, CA, USA, 1964.

[8] M. S. Agbabian, Design of Structures to Resist Nuclear Weapon Effects, ASCE Manual on Engineering Practice, Vol. 42, ASCE, Reston, VA, USA, 1985.

[9] J. H. Prevost, DYNAFLOW: A Nonlinear Transient Finite Element Analysis Program, Princeton University, Princeton, NJ, USA, 1981.

[10] F. Rischbieter and K. Michelmann, Bodenwelle bei Gasexplosionen. Beitrag der induzierten Bodenwelle zur Belastung von KKW-Strukturen, Batelle-Institut, Frankfurt, Germany, 1984.

[11] K. J. Kim and S. E. Blouin, Response of Saturated Porous Nonlinear Materials to Dynamic Loadings, Applied Research Associates Inc., South Royalton, VT, USA, 1984.

[12] H. Werkle and G. Waas, "Analysis of ground motion caused by propagating air pressure waves," Soil Dynamics and Earthquake Engineering, vol. 6, no. 4, pp. 194-202, 1987.

[13] Z. Yang, "Finite element simulation of response of buried shelters to blast loadings," Finite Elements in Analysis and Design, vol. 24, no. 3, pp. 113-132, 1997.

[14] Z. Wang, H. Hao, and Y. Lu, "A three-phase soil model for simulating stress wave propagation due to blast loading," International Journal for Numerical and Analytical Methods in Geomechanics, vol. 28, no. 1, pp. 33-56, 2004.

[15] Y. Lu and Z. Wang, "Characterization of structural effects from above-ground explosion using coupled numerical simulation," Computers & Structures, vol. 84, no. 28, pp. 1729-1742, 2006.

[16] T. Bierer and C. Bode, "A semi-analytical model in time domain for moving loads," Soil Dynamics and Earthquake Engineering, vol. 27, no. 12, pp. 1073-1081, 2007.

[17] S. D. Wilson and E. A. Sibley, "Ground displacements from air-blast loading," Journal of the Soil Mechanics and Foundations Division, vol. 88, no. 6, pp. 1-32, 1962.

[18] M. L. Baron, I. Nelson, and I. Sandler, Investigation of Air Induced Ground Shock Effect Resulting from Various Explosive Sources. Report 2, Influence of Constitutive Models on Ground Motion Predictions, Weidlinger Associates, New York, NY, USA, 1971.

[19] Z. Wang and Y. Lu, "Numerical analysis on dynamic deformation mechanism of soils under blast loading," Soil Dynamics and Earthquake Engineering, vol. 23, no. 8, pp. 705-714, 2003.

[20] S. Wang, L. Shen, F. Maggi, A. El-Zein, and G. D. Nguyen, "Uniaxial compressive behavior of partially saturated granular media under high strain rates," International Journal of Impact Engineering, vol. 102, pp. 156-168, 2017.

[21] W. R. Perret, Ground Motion Studies at High Incident Overpressure, Sandia Corporation, Albuquerque, NM, USA, 1960.

[22] UFC Manual, Structures to Resist the Effects of Accidental Explosions, US DoD, Washington, DC, USA, UFC 3-340-02, 2008.

[23] S. B. Batdorf, "Air-inducedground shock: a one-dimensional theory," in Proceedings of International Symposium on Wave Propagation and Dynamic Properties of Earth Materials, Albuquerque, NM, USA, August 1967, G. Triandafilidis, Ed., pp. 423-432, Aerospace Corp., San Bernardino, CA, USA, 1969.

[24] S. L. Kramer, Geotechnical Earthquake Engineering, Pearson Education, Delhi, India, 1996.

[25] X. Tong and C. Tuan, "Viscoplastic cap model for soils under high strain rate loading," Journal of Geotechnical and Geoenvironmental Engineering, vol. 2, no. 206, pp. 206-214, 2007.

[26] J. An, C. Y. Tuan, B. A. Cheeseman, and G. A. Gazonas, "Simulation of soil behavior under blast loading," International Journal of Geomechanics, vol. 11, no. 4, pp. 323-334, 2011.

[27] P. Y. Hicher and J. F. Shao, Constitutive Modeling of Soils and Rocks, John Wiley & Sons, Hoboken, NJ, USA, 2013.

[28] S. Pathak and G. V. Ramana, "Air-blast induced ground displacement," Procedia Engineering, vol. 173, pp. 555-562, 2017.

[29] N. M. Newmark and J. D. Haltiwanger, Air Force Design Manual, Principles and Practices for Design of Hardened Structures, University of Illinois at Urbana-Champaign, Champaign, IL, USA, 1962.

[30] A. J. Hendron Jr. and H. E. Auld, "Effectof soil properties on the attenuation of airblastnduced ground motions," in Proceedings of International Symposium on Wave Propagation and Dynamic Properties ofEarth Materials, Albuquerque, NM, USA, August 1967, G. Triandafilidis, Ed., pp. 29-47, University of Illinois, Urbana, IL, USA, 1969.

[31] I. Nelson and G. Y. Baladi, "Outrunning ground shock computed with different models," Journal of the Engineering Mechanics Division, vol. 103, no. 3, pp. 377-393, 1977.

Shashank Pathak [ID] and G. V. Ramana

Department of Civil Engineering, Indian Institute of Technology Delhi, New Delhi, India

Correspondence should be addressed to Shashank Pathak; shashankpathaks@gmail.com

Received 19 August 2017; Accepted 5 December 2017; Published 8 March 2018

Academic Editor: Chiara Bedon

Caption: Figure 1: Schematic representation of dilatational stress distribution in ground at time [t.sub.i].

Caption: Figure 2: Schematic diagram for stress-strain relation at a depth "z."

Caption: Figure 3: Scaling of equivalent ground media.

Caption: Figure 4: Measured and equivalent (linear) overpressure time-histories for shot Priscilla at stations (a) P1, (b) P2, (c) P3, and (d) P4 (measured overpressure time-histories [21]).

Caption: Figure 5: P-wave velocity and geomaterial density profile at Frenchman Flat, NTS [21].

Caption: Figure 6: Variation of error in estimated peak displacement using constrained modulus determined from different experimental techniques: (a) T3 and T4 and (b) T1, T2, T5, and T6. T1: deduced from observed ground motion; T2: calculated from seismic velocity; T3: triaxial tests: initial loading; T4: dynamic 1-D compression test: initial loading; T5: triaxial tests: reloading; T6: resonant column test.

Caption: Figure 7: Variation of peak ground displacement with attenuation characteristic length [L.sub.w].

Caption: Figure 8: Effect of (a) attenuation; (b) strain recovery ratio; (c) velocity ratio on estimated ground displacement; (d) velocity ratio on time of occurrence of peak displacement.

Caption: Figure 9: Comparison between measured and estimated peak ground displacements.

Table 1: Displacement solutions for the double-layered media. Case Explanation Displacement equation 1 When all loading and unloading fronts are in layer-1 ([t.sub.i] [less than or equal to] H/[V.sub.p[). When some loading fronts have crossed layer-1 and remaining loading fronts [mathematical along with all unloading expression not fronts are in layer-1 (H/[V.sub.p] reproducible] < [t.sub.i] < fH/[V.sub.p]). Here, the first task is to determine those loading fronts which have crossed layer-1. [mathematical This can be obtained by equating expression not 2 loading front velocity reproducible] multiplied with travel time with the depth of layer. The loading front velocity of the kth front can be written using (3) as [V.sub.p]/{ (f - 1)([t.sub.k]/ [t.sub.r]) + 1}, and travel time is ([t.sub.i] - [t.sub.k]). Thus, the last loading front to reach at depth H by the time [t.sub.i] would be given by [t.sub.kH] = [t.sub.r] ([V.sub.p] [t.sub.i] - H)/{H(f - 1) + [V.sub.p] [t.sub.r]}. When all loading fronts are in layer-2 and some unloading fronts have crossed layer-1 (fH/ [V.sub.p] < [t.sub.i]). 3 Similar to case 2, the last unloading front to reach at depth H by the time [t.sub.i] would be given by [t.sub.kH] - [t.sub.i] [mathematical - fH/[V.sub.p]. expression not reproducible] Table 2: Details of recording stations. Station Range Peak Positive phase designation (m) overpressure duration (sec) (kPa) P1 198 1863 0.196 P2 259 1290 0.179 P3 320 828 0.307 P4 411 408 0.442 Section Positive Equivalent Scale factor designation phase impulse duration (SF) (kPa-sec) (sec) P1 88.32 0.0948 2.07 P2 62.18 0.0964 1.86 P3 50.54 0.1221 2.51 P4 48.34 0.2371 1.86 Table 3: Constrained modulus of playa silt at Frenchman Flat [4]. Method Notation Constrained modulus (MPa) For For depths depths less more than 18 m than 18 m Deduced from observed ground motion T1 96.6 552 Calculated from seismic velocity T2 242 1035 Triaxial tests: initial loading T3 55.2 103.5 Dynamic 1-D compression test: initial loading T4 69 138 Triaxial tests: reloading T5 276 448.5 Resonant column test T6 276 759 Table 4: Scaling of attenuation characteristic length ([L.sub.w]). Station W [P.sub.o] [V.sub.p] [L.sub.w] Scaled designation (kt) (kPa) (m/s) (m) [I.sub.w] (m) P1 1863 658.69 48.58 23.47 P2 37 1290 658.69 58.38 31.44 P3 828 658.69 72.88 28.98 P4 408 658.69 103.84 55.70 Table 5: Parametric studies carried out on the proposed closed-form solution. Parameter Parametric study [M.sub.1] Parameters [M.sub.1] and [M.sub.2] based on and six different methods are adopted from [M.sub.2] Table 3. Other parameters kept constant at r = 0.6; f = 2; [L.sub.w] adopted from Table 4. Variation of absolute percentage errors in estimated peak ground displacements is plotted against peak overpressures in Figures 6(a) and 6(b). [L.sub.w] [L.sub.w] is varied from 10m to 250m. Other parameters kept fixed at [M.sub.1] and [M.sub.2] adopted from Table 6; f = 2; r = 0.6. Estimated peak displacements are plotted against [L.sub.w] as shown in Figure 7. Errors in estimated peak displacements for the attenuating medium and non-attenuating medium are also plotted against peak overpressure in Figure 8(a). r Two cases are considered: (i) full strain recovery r = 1 and (ii) partial strain recovery r = 0.6. Other parameters fixed at [M.sub.1] and [M.sub.2] adopted from Table 6; f = 2; [L.sub.w] adopted from Table 4. Errors in estimated peak displacement for the two cases are plotted as shown in Figure 8(b). f f is varied from 1.0 to 2.2. Other parameters fixed at [M.sub.1] and [M.sub.2] adopted from Table 6; r = 0.6; [L.sub.w] adopted from Table 4. Estimated peak displacements are plotted against f (Figure 8(c)). Error in occurrence time of peak displacement are also plotted against f (Figure 8(d)). Observation [M.sub.1] Peak ground displacements are very and sensitive to the constrained modulus [M.sub.2] value. Average coefficient of variation in peak displacement estimates = 83%. Estimates are close to measured values if higher modulus values are used for smaller overpressures. [L.sub.w] As [L.sub.w] increases (or [alpha] decreases), peak displacement increases. Beyond [L.sub.w] of 250m, the peak ground displacement does not increase, and [L.sub.w] [greater than or equal to] 250 is considered as the non-attenuating medium. r Assumption of full strain recovery gives less errors as compared to partial strain recovery under low overpressures. f Under high overpressure (P1 and P2), error in occurrence time of peak displacement reduces significantly with increasing f. Error increases marginally under lower overpressures as the velocity ratio is close to 1 under lower stress (in (3) when [t.sub.k] [right arrow] 0). Recommended value [M.sub.1] Observed results are consistent with the and observations of Wilson and Sibley [17]. [M.sub.2] (1) Shallow ground layers are likely to have modulus values determined by unconfined or triaxial compression test, and deep ground layers are likely to have modulus values computed from seismic velocity test or resonant column test.+e justification to this variation in selection of modulus values can be attributed to the small strains associated with deeper layers and higher strains at shallow depths. (2) Constrained modulus increases with decreasing ratio of applied stress to overburden. For higher overpressures, the overstress ratio would be higher and therefore modulus value will be lower compared to lower overpressures. An optimal choice of constrained modulus values is adopted as shown in Table 6. The computer displacements are found to be in good agreement with measured displacements (Table 6). [L.sub.w] Attenuation has to be taken into account under low overpressures, and it can be ignored under high overpressures. r Under low overpressure, ground is not stressed beyond its elastic limit, and hence full strain recovery is a better representation of actual conditions in low overpressure zones. A lower strain recovery causes higher permanent deformations and increases peak ground displacement compared to elastic case (i.e., unit strain recovery ratio). Under higher overpressures, the ground is stressed beyond its elastic limit and the assumption of partial strain recovery is recommended. f f has insignificant effect on magnitude of estimated peak displacements. However, as the velocity ratio increases, the rise- time of overstress pulse with depth also increases and affects the occurrence time of the peak displacement. Table 6: Suggested combination of modulus values and corresponding estimated peak displacements. Station [M.sub.1] (MPa) [M.sub.2] (MPa) Estimated designation (depths less (depths more peak than 18 m) than 18 m) displacement (mm) 55.2 552 P1 triaxial tests: deduced from 235.59 initial loading observed ground motion 69 759 P2 dynamic 1-D 147.68 compression resonant column test: initial test loading 69 759 P3 dynamic 1-D resonant column 74.79 Compression test test: initial loading 96.6 1035 P4 deduced from calculated from 38.49 Observed seismic velocity ground motion Experimental % records (mm) error P1 250.95 6.12 P2 157.56 6.27 P3 68.75 8.79 P4 33.95 13.37

Printer friendly Cite/link Email Feedback | |

Title Annotation: | Research Article |
---|---|

Author: | Pathak, Shashank; Ramana, G. V. |

Publication: | Advances in Civil Engineering |

Date: | Jan 1, 2018 |

Words: | 5913 |

Previous Article: | Stability Analysis and Confidence Level Evaluation of Backfill Mining under High and Steep Rock Slopes. |

Next Article: | Shield-Roof Interaction in Longwall Panels: Insights from Field Data and Their Application to Ground Control. |

Topics: |