Ground shock attenuation on a buried cylindrical structure by a barrier

This paper presents results from studies on the influence of an attenuation barrier on soil stresses and pressures acting on a buried silo, caused by underground explosions at different distances from the structure. The attenuation barrier was made of PVC tubes that were placed between the explosive charges and the RC silo structure. Soil stress gages and pressure transducers were used to measure these data in the soil and on the structure, respectively. The influences of several parameters were analyzed for tests with and without an attenuation barrier between the silo and the explosive sources.


Introduction
Only a few studies have been published on protecting underground structures from ground shock induced by buried explosions [6,9]. Buried structures, or buried parts of above-ground structures, may be exposed to such ground shock during wartime, terrorist activities and tunneling during subway construction with drill-and-blast technique activities, as well as drill shaft for transmission line pole foundations activities. Accidental explosions of subterranean gas pipelines or tanks could induce similar effects. Protective barriers (screens) may contribute to decrease such effects to acceptable levels. Following types of barriers can be considered for such applications [14,17,18]. a) Barriers with a low acoustic impedance (LAI), e.g., polyester, styrofoam, etc. b) Stiff barriers e.g., reinforced concrete (RC), c) Barriers made of particular media. d) Gas cushions barriers One must differentiate between the following two basic types of loading conditions that can occur during an underground explosion in order to evaluate particular barriers [17].
1. Shock loading generated by the stress wave from the blast, as it passes through the soil. 2. Inertial loading resulting from the physical movement of the block of soil located between the detonating charge and the structure. The PVC barrier is characterized by its considerable shock load attenuation ability. A stiff barrier, in turn, can attenuate inertial loading substantially. Combinations of various types of barriers (e.g., layered polyester-particulate media-polyester) have also been utilized. The attenuation results of such barriers may be found in the work of Williams et al. [17]. To this end, they used a simplified formula to define the degree of attenuation (transmission coefficient) [8]: Where: I T is the intensity of the waves transmitted through the barrier, and I t is the intensity of incident waves on the barrier. The intensity of waves is defined by the relationship I = σ 2 max /ρ · c in which σ max is the peak pressure, ρ is the medium density, and c is the wave propagation velocity. They showed that the above noted barrier types contribute to a decrease in the maximum stress levels. The greatest attenuation was attained when particulate and combined barriers were used. The smallest attenuation occurred when reinforced concrete barriers were used. The value of the attenuation coefficient C t is shown in Table 1.
Woods [18], in turn, investigated the influence of an air-gap (i.e., trench type) barrier on the reduction of vibrations. The first scientific investigations on the barrier effect of trench type at large scale were reported by Barkan [2]. Woods at al. [19] studied also the effectiveness of barrier made of rows of discrete cylindrical voids (piles). According to Woods [18] (see [7]) a barrier can be considered as either active isolation or passive isolation. Active isolation is in the case where the barrier completely or partially surrounds the source of vibration continuously (for example circular open trench, Fig. 1a). Passive isolation appears in the case of trench segments (Fig. 1b), and discrete cylindrical voids or holes (piles, Fig. 2b). Woods concluded from the field investigations on passive isolation (e.g., an open trench barrier of depth H, length L, width W , Fig. 2a) that the reduction in the applied vertical vibrations induced through an exciter footings resting at the ground level may be summarized, as follows: For a satisfactory passive isolation (for R = λ r to about 7 λ r ), the minimum trench depth H should be between about 1.2 and 1.5 λ r and H/λ r should be about 1.33, where λ r is the wavelength of the Rayleigh waves. One should point out that most of the vibratory energy affecting nearby structures in that case is carried by Rayleigh (surface) waves traveling from the source of vibration. Woods et al. [19] also carried out model tests of passive isolation using both single and double rows of cylindrical holes as an energy barrier. These tests were performed in a box filled with a fine-sand medium (Fig. 2b). Based on these test results they suggested that a row of cylindrical holes may act as an isolation barrier if and, where, D is the diameter of the cylindrical void, and S n is the net space for the energy to penetrate between two consecutive void obstacles. The literature comprises many theoretical investigations regarding the efficiency of vibration isolation barriers, using the Finite Element Method and Boundary Element Method [1,3,4,10]. For a discussion on these investigations, reference is made to Haupt [11] and Massarsch [14].
This paper presents results reduction of soil pressures acting the interface between a RC silo structure and soil as well as of stresses within soil by PVC tubes (passive isolation) placed between buried explosive charges and the buried structure.

Tests program and site
Five PVC tubes with a 100 mm. diameter and 2.0 m. length were placed in the soil at 1.0 m. from external surface of a RC cylindrical silo. The empty tubes were located in the three layered soil so that they shielded (protected) the silo structure (D = 1.78 m -external diameter and H = 3.40m -height) from the ground shock (Fig. 3). The influence of the attenuation barrier on the reduction of stresses σ h inside the soil, measured by soil stress gages and the horizontal soil pressure p h acting on the external surface of the silo wall measured by pressure gages, were the subject of study. The influence of distance R and equivalent mass W TNT of charges in the tests of these quantities were analysed. The depths of charges installation h were the same in all tests. The accelerations g of top point of silo were also measured. The full program of six tests is presented in Fig. 4. The field test scheme with the buried concrete silo structure, arrangements of pressure gages fixed to the wall on zdepths and soil stress gages and data acquisition are given on Fig. 5. Figure 5 also shows the position PVC tubes screen as well as charges and the sequence of their explosion denoted by numbers 1 ÷ 6. It should be noted that the soil (Fig. 3) was undisturbed only in the first test, and its properties are given in [13].
The following parameters are introduced to enable using the presented results and their analyses for determining the blast-induced pressure acting on the silo wall for other sizes but similar structures in which all dimensionless similarity parameters π 1 , π 2 , π 3 , . . . , π 13 from Eq. (1) are satisfied [13]: ψ(π 1 , π 2 , π 3 , . . . , π 13 ) = p h γsD -represents the horizontal pressure p h acting on the external surface of silo wall, -π 1 = R/D -represents the distance R from the charge to the front of silo wall, Fig. 2. Passive isolation tests, the plan view of (a) field site, and (b) the model half-space in a box of layouts in [18,19].
π 2 = H/D -represent silo geometry, -π 3 = h/D -represents the burial depth of the charge, -π 4 = z/D -represents the depth at which the pressure is measured, -π 6 = WTNT γSD 3 -is the scaled mass of the charge where γ s =ρ s g -specific soil density, ρ d -soil particle density, z -depth at which pressure is measured (see Fig. 5).

Soil stresses
The stress measurements in the horizontal direction within the soil were made by means of a disk shape soil stress gages with a flexible epoxy external ring (Fig. 6) for reducing the gage cross-sensitivity and the overall unit weight [5,12]. Soil stress gages and charges were dropped into previously prepared 1.8 m deep bore holes. The  holes were filled with the soil and compacted sensor placement. Attention was given during the sensor installation to precision of hole backfill and compaction around the stress gages to reproduce the in situ physical properties of virgin soil [16].  Figure 7 shows three soil stresses-time histories measured by Gage No. 0, as a function of the distance from the detonated charge (charge placement depth 1.8 m) with the barrier present. It should be emphasized that first charge was detonated at the farthest location from the soil stress gage, and the charges were gradually placed closer in subsequent tests. Therefore, the soil was undisturbed (virgin) only during the first test i.e. for the largest distance. The wave speed was the largest for the undisturbed soil, and it decreased in subsequent explosions, as observed from Fig. 7 where the arrival wave times from the explosion sources to soil stress gages are denoted. The calculated wave speed (distance divided by arrival time) for the first, second, and third explosions are 38.7 m/sec, 24.5 m/sec and 23.3 m/sec, respectively.

Pressures acting on an embedded silo structure in soil
Figures 10 to 20 present the normal pressure-time histories induced on the structure, measured by pressure gages installed on the wall surface of the front, rear, and side generatrixes and, for various charge-to-structure distances, with and without an attenuation barrier (screen). Figure 10 shows a comparison between the pressure on the wall recorded by gage No. 1, without a barrier in the path of the propagating shock wave with that for the case with a barrier placed 1.0 m before the front of the silo wall. The barrier is marked on the figures by a vertical line to the left of the silo. This comparison shows very clearly that for a detonation at a distance of 7.0 m the attenuation barrier reduced the magnitude of the recorded peak interface pressure by more that 75%, from about 6.1 kPa to about 1.5 kPa ([6.1-1.5]/6.1] × 100%).
An assessment of the barrier's influence may be obtained by an attenuation coefficient C t , as shown in Eq. (4): In which, p hb is the interface pressure on the structure wall with the barrier (attenuated pressure), and p hf is the interface pressure on the wall without a barrier (free field or not-attenuated pressure).     (Fig. 20), 0.33, 0.14 and 0.18, respectively. Figure 18 presents two pressure-time histories recorded by gage No. 6 (z/D = 1.01) placed on the rear structure surface, with and without an attenuation barrier during explosion of a 1.2 kg charge at a distance of 5.7 m. In this case, the attenuation is small, in spite of that this part of structure is screened by the barrier and soil is no so much destructed by only one previous detonation; the attenuated pressure was 4.26 kPa vs. the unattenuated pressure of     which is below of the barrier. Significantly smaller pressure attenuation were obtained in the case screened gage placed on depth z = 1.80 m.    Figure 24 shows the maximum pressure parameter ψ with and without attenuation along the height vs. the explosion distance parameter π 1 . The values of π 1 and the corresponding ranges for R were, as follows: 3.93 for 7.0 m, 3.20 for 5.70 m, 2.64 for 4.7 m, and 2.07 for 3.7 m. This comparison of the horizontal pressure coefficients with and without attenuation indicates very clearly that the pressure decreased during attenuation with all other conditions unchanged. But one should remember that the soil conditions in successive detonations from 1 to 6 deteriorated, and resulted in higher pressures on the silo wall than in the case of undisturbed soil.

Analysis of interface pressure on a wall of a embedded silo with an attenuation barrier
As may be seen in Fig. 25, the smallest attenuation value corresponds to the case when the explosion occurs at the smallest distance from the barrier (π 1 = R/D = 2, 07). Also one can notice two different attenuation trends: one for π 1 of 3.93 and 3.20, and the second for π 1 of 2.67, and 2.07. In the first case i.e. for longer charge-to-silo distance it shows an equal barrier influence along the silo height with no changed soil properties, and in the second   case i.e. for smaller charge-to-silo distance one can observe the best attenuation on the silo bottom and a poor barrier influence on the silo top, especially for π 1 = 2.07 with some influence of soil destruction caused by successive

explosions.
The relationships between the structure accelerations and the various parameters were also studied, as shown in Fig. 26 (unattenuated), 27 (attenuated), and 28 (comparisons between the two). Figures 26 and 27 show the relationship between acceleration and detonation distance with and without the barrier present, respectively. One can note that the structure's accelerations is larger for shorter charge-to-silo distances. Of course, structural accelerations values are larger without a barrier. One may conclude that for longer distances the acceleration is unaffected by a barrier. It can be clearly noted from Fig. 28 that the barrier affects the acceleration for charge distance of less than 5.7 m, while it has no effect for larger charge distances.

Conclusions
The following conclusions were obtained based on the findings from this pilot study: 1) The effectiveness of the attenuation barrier in regard to pressure reduction increased with the increased distance between the explosion source and the silo, when the structure-to-barrier distance was unchanged.
2) The barrier is ineffective in shielding the upper part of the structure when the explosion occurs very close to the barrier. However, the barrier can shield the structure only if its strength is sufficient, and it will not fail under the incoming ground shock But one should take under consideration the destruction of the soil by previous explosions.
3) It was noted that very high pressures (i.e., small attenuation) were registered by the gages placed on the side (C t = 0.93) and rear (C t = 0.89) of the silo surface at mid height (z/D = 1.01). 4) The structural acceleration increased with increased charge weight and detonation depth. 5) Additional studies are required to obtain more complete explanations of these phenomena. For example, to consider the effects of different soil conditions (e.g., others soil profiles), a broader range of structural and barrier dimensions, relative structure-to-barrier locations, charge sizes, and detonation depths and charge-tostructure distances. Because the detonations change the native soil properties, the soil should be restored in each successive tests.