Internal Explosions in Embedded Concrete Pipes Pamela Bonalumi1,a, Matteo Colombo1,b and Marco di Prisco1,c 1
Department of Structural Engineering, Politecnico di Milano, Milan, Italy
a
[email protected],
[email protected],
[email protected]
Keywords: internal explosions, blast tests, embedded pipes, concrete structures, soil-structure interaction.
Abstract. Blast tests on a full-scale concrete pipe embedded in soft soil were carried out to evaluate the behavior of the soil-structure system under the internal detonation of high-energy solid explosives. Two different stages were considered: the first one focused on the detonation of a low entity charge within the pipe, to maintain the concrete in the elastic regime, and the second one concerned with adopting larger quantities of explosive to produce a local damage to the structure. Cylindrical charges ranging from few grams to hundreds of grams of a high-energy solid explosive were investigated and different tests were performed for each quantity by inserting the explosive charge in a card cylinder hanged up in the middle of the pipe central segment by means of three thin plastic wires. The following quantities were measured in different sections along the pipe: side-on and reflected pressure-time histories at the inner surface of the structure, pipe radial acceleration; peak particle acceleration of the surrounding soil by means of accelerometers placed at different distances and depths from the section where the explosion occurred. The experimental results obtained during the performed blast tests are so analyzed to understand the soil-structure system behavior under such fast transient dynamic phenomena. Introduction The analysis of blast-resistant structures has become an active topic in the civilian field from the early nineties, due to a series of worldwide terrorist events such as the truck bomb explosion in the World Trade Center in New York City in 1993 and the bomb explosions at the financial centers of London in 1994. However, not many established standards or practices on the design of civilian blast-resistant structures have been developed, also because the methodologies and techniques acquired for the protection of military facilities were often denied to the civilian sector [1]. The more recent attacks to the World Trade Centre in New York and to the London tube have caused considerable concern on the protection of structural integrity to assure safety conditions to occupants. Different research and tests have been sponsored by U.S. government agencies to provide manual on protective structures [2, 3], but the focus was mainly on above ground structure subjected to external explosion or box structure types under internal explosion, neglecting any references to internal explosions in underground structures such as tunnels. Besides, experimental studies related to any particular combination of structure, soil and loading are really scarce, especially because full-scale experiments are really expensive and not so easy to be performed. Due to the lack of experimental data, the blast tests here presented were carried out to investigate the behavior of a full-scale soil-structure system, moving a first step towards the collection of useful experimental data capable of providing a first sight on the relations existing between the blast wave propagation within an embedded structure, the consequent structural response and the shock wave effects on the surrounding soft soil, which represents a more disadvantageous condition than pipe embedded in rocks [4]. Blast tests set-up The experimental investigation concerned a series of blast tests on a full-scale embedded concrete pipe, subjected to a single internal detonation due to a solid high-energy explosive.
The tests were carried out at the training campus of the Lombardy Civil Protection and Fire Brigade, placed in Bovisio Masciago (MB) on a plain concrete pipe 26 m long and 85 mm thick, with an inner diameter of 1.0 m, embedded in soft soil at a depth of about 2.30 m. The pipe is made of different precast segments, characterized by a length of about 1m each, and is connected at each end side to a shaft (Figure 1).
a)
b)
Fig. 1: a) Area interested by the blast tests in Bovisio Masciago; b) Embedded plain concrete pipe. Test outline. The blast tests consisted of two main stages, as detailed in Table 1. Since blast tests were carried out on a structure already existing in situ, a series of laboratory and in situ tests were preliminary performed to characterize the mechanical properties of the concrete constituting the tube and the soil surrounding the pipe. First step. The first step focused on the detonation of low entity charges (10 g and 12 g) within the pipe, to maintain the concrete in the elastic regime in order to avoid irreversible damages to the structure and to assure the repeatability needed to provide statistical validity to the experimental results. In this first step, 12 different tests were carried out for 10 g charges, performing 6 tests by placing the charge in the middle of the tube central segment (Section C) and other 6 tests by locating the charge in the first instrumented section near the structural joint (Section J), to investigate the possible influence of the structural joint. Other 6 tests were carried out detonating 12 g charges in the middle of the tube central segment, to compare the effect of different charge weights. Second step. The second stage concerned the detonation of larger quantities of explosive (120 g and 120 g) to produce a local damage to the structure. It will not be treated in the following. Explosive charges. In the tests here presented dynamite ERGODYN 35E was adopted as highenergy solid explosive, due to its physical properties (Table 1) and its capability of fulfilling security requirements. To assure a good compromise between theoretically symmetric wave propagation and handiness of the explosive case, a cylindrical charge was adopted. For the smallest charges (10 g and 12 g), the solid explosive was inserted in a cardboard cylinder and confined by two pieces of plasters, placed ahead and beyond the true explosive, in order to assure a frank detonation of such small charges. Before putting the plaster, a 1 g booster was inserted in the card cylinder and put in contact with one end of the explosive. The cylinder manufactured in such a way was hanged up in the middle of the pipe section (Section C or Section J) by means of three thin plastic wires, not to interfere with the blast waves propagating from the charge. Such position of the charge was chosen not only to assure an initial symmetric condition, but also to avoid contact problem between the structure and the explosive, such as large localized plastic deformations of the tube and fragments.
An electric detonator was adopted to fire the charge and, despite it triggered both the firing and the data acquisition system (DAQ), the charge ignition and the data acquisition were not perfectly synchronized with the firing because of the time delay caused by the electric signal traveling down the cables. Dynamite ERGOGYN 35 E Density Detonation velocity Detonation energy: - Shock energy - Gas energy - Total energy Detonation pressure Gas volume at 0°C/atm
1400 kg/m3 5900 m/s 2300 MJ/kg 2000 MJ/kg 4300 MJ/kg 20000 MPa 865 l/kg
Table 1: Explosive properties.
Fig. 2: Explosive charges.
Instrumentation. The concrete tube was properly instrumented in 5 different sections along the tube with: 7 reflected pressure transducers, 1 incident pressure transducers, 8 uni-axial accelerometers and 10 strain gauges (just during the second step), as shown in Figures 3. Moreover 4 biaxial accelerometers were placed within the surrounding soil at different distances and depths from the central section of the pipe. Thus, during the blast tests the following physical parameters were measured: side-on pressure-time history within the structure and reflected pressure-time histories on the internal surface of the pipe by means of pressure transducers; radial and longitudinal accelerations of the structure by means of uni-axial accelerometers in contact with the outer surface of the tunnel; the pipe hoop strains by means of strain-gauges and the acceleration of the surrounding soil by means of bi-axial accelerometers placed in the ground. The data acquisition system was a 56 channels parallel system with maximum sampling rate of 3MS/s/ch and 14 bit resolution. All channels were acquired at a sampling rate of 1 MS/s.
Fig. 3: Blast tests set-up of Step 1: a) longitudinal section; b) transversal section. Blast tests results The experimental data recorded during the different blast tests performed on the embedded concrete tube are here presented in terms of pressure, acceleration and strain-time histories. It is worth noting that all time histories refer to an arbitrary time instant chosen as initial time, due to the fact that the data acquisition process starts with the trigger event, which slightly differs
from the effective time of detonation, as previously explained. All data were filtered to minimize artifacts and noise from the incoming signals, as described in detail for each kind of signal. The position and the identification number of the transducers that the curves refer to are reported in each graph, with reference to the experimental set-up shown in Figure 3. Different aspects can be outlined from the collected experimental data, such as the repeatability obtained by detonating small charges of the gram-order, which can be particularly influenced by the way of packaging and the process of internal detonation and the effective symmetry of the phenomena. In the following, just results concerning the detonation of 10 g charges near the structural joint (Section J) will be presented (see [5] for further results). All the following considerations, but repeatability, refer to the resulting average curve recorded for the 6 different tests performed by each sensor. Side-on and reflected pressure. The pressure-time histories recorded by the incident pressure transducer, located in the middle of the pipe section at about 4 m from the explosion point, and by the 7 reflected pressure transducers, screwed in the thickness of the pipe in 5 different sections, are presented in the following. All signals were filtered by an acausal passband Butterworth filter of the third order, between frequencies of 100 Hz and 15000 Hz to cut the electric current noise, characterized by a frequency of about 50 Hz, and really high spurious frequencies, without loosing important components.
a)
b)
Fig. 4 : Repeatability of pressure-time histories: a) reflected pressure on the pipe inner surface in the explosion section; b) incident pressure in the centre of the pipe section at 4 m from the explosion. From the analysis of side-on and reflected pressure time-histories in Fig. 4, it can be noticed that the blast load that impinges on the structure show a really good repeatability not only far from the charge, but also in the explosion section, at a standoff distance R of about 0.5 m. The reflected pressure-time history, presented up to 8 ms to underline the broadband of the signals, shows the strong impulsive nature of the blast load, characterized by a strong first peak followed by negligible reflections, acting on the explosion section. The side-on pressure instead is characterized by a wave train after the arrival of the incident blast front, characterized by a maximum pressure value comparable with the first peak pressure. The pressure-time histories recorded by the couple of symmetric transducers located in the pipe section where the explosion occurs (Fig. 5a) highlight the good symmetry of the blast load impinging on the structure, not only in terms of the first reflected peak but also concerning the following. Looking at the longitudinal propagation of the blast wave in Fig. 5b, it can be observed the exponential decay of the pressure peak value, which is really rapid close to the explosion point and strongly decreases with distance from the charge. The first peak pressure recorded in Section 3, at about 1 m from the explosion point (equal to two times the internal diameter of the pipe), is equal to
the peak pressure of the second reflection recorded in the explosion section (at about 0.5 m form the charge).
a)
b)
Fig. 5: a) Symmetry of the reflected pressure-time histories on the explosion section; b) Longitudinal propagation of the reflect pressure-time histories. Structural radial accelerations. The acceleration-time histories recorded by the 8 radial accelerometers located on the outer surface of the pipe in 5 different sections are presented in the following. In this case, all signals were filtered by an acausal passband Butterworth filter of the third order, between frequencies of 100 Hz and 10000 Hz, to cut the noise due to electric current, characterized by a frequency of about 50 Hz, and to high frequencies above which the accelerometer frequency response ceases to be within the flat region.
a) a)
b) b)
Fig. 6: Repeatability of radial acceleration: a) explosion section; b) at 4 m from the explosion. The structural response in terms of radial accelerations presents a good repeatability both in the segment where the explosion occurs and in the farthest segment (Section 5), confirming the absence of damage and the elastic behavior of the structure (Figure 6). The section where the explosion occurs undergoes really high peak values of the radial acceleration and presents a clear impulsive behavior, related to the nature of the blast load acting on it. On the contrary, an oscillatory behavior characterized by smaller peak value of about one order of magnitude is pointed in Section 5, at about 4 m from the explosion. Looking at the frequency content of the radial accelerations recorded recoded at different distances from the source (Fig. 7), it can be observed that the signal presents a broadband behavior up to 450-500 Hz and a significant frequency content also for higher
frequencies around 1000 Hz close to the explosion, which rapidly decrease with the distance form the source. At lager distances from the charge, just the frequency content around the lowest frequencies of 450-500 Hz remain in the signal.
Fig. 7: Fourier Amplitude Spectrum of the radial acceleration as function of distance from explosion. Conclusions The behavior of a plain concrete pipe embedded in soil and subjected to the internal detonation of a solid high-energy explosive, placed in the centre of the structure middle section, has been investigated. Despite the small charges of the gram-order adopted, there is a quite good repeatability and symmetry of the blast load that impinges on the internal surface of the structure after the detonation of a cylindrical solid high-energy explosive within a pipe. The local effects, such as the positioning of the charge, the packaging of the charge and the internal detonation process within the charge seem not to significantly affect the load acting on the structure at the small scaled distances here investigated. A good repeatability of the structural response in terms of accelerations has also been observed and a strongly different behavior of the section has been highlighted for the pipe sections at different distances from the explosion, both in the time and in the frequency domain. Acknowledgment The authors gratefully acknowledged the financial support of the European INTERREG IT/CH 2006_2013 project ACCIDENT ID 7629770, Measure 2.2, which funded the presented research. The authors would like to thank Admiral Roberto Vassale for the precious work and the Municipality of Bovisio Masciago for the collaboration. References [1] National Research Council. Protecting buildings from bomb damage: transfer of blast effects mitigation technologies from military to civilian applications, National Academic Press Edition, 1995. [2] DOE/TIC-11268. A Manual for the Prediction of Blast and Fragment Loadings on Structures. U.S. Department of Energy, 1992. [3] UFC 3-340-02. Unified Facilities Criteria (UFC) - Structures to resist the effects of accidental explosions. HQUSACE, NAVFAC and AFCESA, December 2008. [4] H. Liu. Dynamic analysis of subway structures under blast loading. Geotechnical and Geological Engineering, 27(6): 699–711, 2009. [5] P. Bonalumi. Soil-structure interaction under impulsive loading: internal explosions in embedded pipes. PhD Thesis, 2011.
