Geotechnical Aspects of Explosive Compaction

Explosive Compaction (EC) is the groundmodification technique whereby the energy released from setting off explosives in subsoil inducing artificial earthquake effects, which compact the soil layers. The efficiency of EC predominantly depends on the soil profile, grain size distribution, initial status, and the intensity of energy applied to the soil. In this paper, in order to investigate the geotechnical aspects, which play an important role in performance of EC, a database has been compiled from thirteen-field tests or construction sites around the world, where EC has been successfully applied for modifying soil. This research focuses on evaluation of grain size distribution and initial stability status of deposits besides changes of soil penetration resistance due to EC. Results indicated suitable EC performance for unstable and liquefiable deposits having particle sizes ranging from gravel to silty sand with less than 40% silt content and less than 10% clay content. However, EC is most effective in fine-to-medium sands with a fine content less than 5% and hydraulically deposited with initial relative density ranging from 30% to 60%. Moreover, it has been observed that EC can be an effective method to improve the density, stability, and resistance of the target soils.


Introduction
Depending on the importance and load level of a structure, ground improvement may be a viable alternative to expensive foundations in conditions involving weak and problematic soil deposits.The common options for improvement of weak soil layers at depth are as follows: vibro methods, Deep Soil Mixing (DSM), Deep Dynamic Compaction (DDC), jet grouting, and Explosive Compaction (EC).Modification of loose, saturated deposits at depth using methods such as Deep Soil Mixing and jet grouting is costly.Dynamic compaction involves dropping of heavy weights to improve soil and is effective to a depth of 15 m, although its efficiency decreases with decreasing soil permeability.Shakeran and Eslami [1] reported that the use of vibrators, for improvement of loose granular soils at depth, have three disadvantages: (a) they require special equipment, (b) their efficiency decreases with increasing silt and clay content, and (c) their application is limited to maximum depth of 30 m.
Compaction of the site of Franklin Fall Dam in New Hampshire in late 1930s was the first successful application of the EC method.Soon after application of EC method in Franklin Fall Dam, the effectiveness of the method was approved and subsequently the method was successfully applied in the compaction of hydraulic deposits of Cape Cod Canal in Texas and Almond Dam in New York [2,3].
In 1970s and 1980s, most of the research pertaining to EC method was carried out by the US army and was concentrated on the study of how blast effects cause liquefaction [4].During those years, examples of projects carried out included (i) compaction of existing deposits beside slopes and structures adjacent to residential areas, (ii) densification of deposits associated with waste pools, and (iii) coastal constructions and execution of the EC in the oil-drilling platform Molikpaq I.
In the 1980s, the capability of the EC technique as a deep soil improvement method was demonstrated when the method was applied to eliminate earthquake liquefaction hazard and improve a 15 m thick deposit of loose, saturated, and unstable alluvial soil existing at a depth of 30 to 45 m from below ground level in Jebba dam project.Later in 1990s and 2000s, the EC method was used on projects relating to embankments of bridges and dams, waste pools, and liquefaction assessment tests [5].Densification of cohesionless soils occurs due to a few mechanisms imposed by the blasting, including compression, volumetric strains, and shearing.Moreover, excess pore water pressure is generated which may induce liquefaction and resettlement of the soil particles into a denser configuration [3].After the first phase of EC and dissipation of residual pore pressure, subsequent phase of blasting induces more settlement in the soil.Amount of final settlement depends on initial stiffness and relative density and blast operation.In other words, first phase of blasting breaks all bonds (due to cementation or aging) between soil particles and other phases increase settlement.Saturated sands are more suitable for blast densification than dry sand because it is a more efficient medium for shock pulse transmission and residual pore water pressures, reducing effective stresses.Figure 1, illustrates wave propagation and the effect of explosion on ground subsidence, soil liquefaction, and surface settlement.
In order to transmit sufficient energy to the soil mass and thereby achieve a suitable soil compaction in the EC method, it is necessary to (a) ensure appropriate arrangement of blast holes on plan including pattern (square or triangular grid) and spacing and (b) select the correct amount of explosives and distribution of them along the thickness of the target layer.

EC, Performance Mechanism, and Effective Parameters
After the first phase of blasting and dissipation of the major portion of excess pore water pressure, repeating the blasts in the subsequent phases will cause additional compaction of soil.After each explosion phase, soil will be more compacted, but soil settlement as well as compaction rate will decrease and after two or three phases of explosion, large settlement will not occur anymore [6].Final compaction will depend on the stiffness and initial density of the soil in addition to the method of setting off the explosives.The first phase of blasting will decompose all the soil grain bonds associated with cementation or aging, whereas the subsequent phases will compact the soil further.Clay particles in sand will decrease the soil drainage and thereby reduce the efficiency of EC.Narin van Court [7] suggested that (a) the minimum CPT cone resistance  t for a soil to be satisfactorily compacted by the EC method is 10 MPa, (b) medium density compaction is achieved in soils with  t less than 15 MPa, and (c) blasting causes loosening in soil deposits where  t exceeds 20 MPa.
Effective parameters in design of EC fall in six categories are as follows: (i) charge weight in each hole, (ii) depth of charges in the hole, (iii) scattering pattern of charges in height, (iv) distance between the blast holes, (v) phasing and the number of blast stages, and (vi) sequence of blast holes and explosive scheduling in each phase.Determination of these design parameters should be based on the final optimum results.Figure 2 illustrates a real plan and profile of blast holes arrangement along with the charging locations and sequences of explosion [8].
The most important results of EC include increasing relative density of the soil (  or   ), increasing soil resistance to cone penetration, and suppression of volume changes that cause settlement of the target.Increase in   of the soil is because of arrangement of the grains in a denser state after EC.As a rule, increase in   after blast densification is a factor of fine contents in the sand and initial relative density.EC can increase in the compactness of loose sands having relative density in the range of 20% to 30% up to 60% to 70%.Comparison of the results of penetration tests such as CPT and SPT before and after blasting indicates that EC can be used to increase the cone resistance and thus the bearing capacity of the soil.Figure 3(a) presents sample CPT records before and after EC [9].The increase of the soil resistance to cone penetration does not occur immediately after blasting but occurs over a period.Figure 3(b) illustrates the increasing PT tip resistant with time after using some soil improvement methods [10].With increasing time after blasting, bonds decomposed by the shock effects are formed again resulting in enhancement of the soil strength.This process may last from several days to even several months [11].

Data Bank and Case Histories
For evaluation of the range of improved deposits using the EC technique, 13 sites were selected from various countries, including countries, namely, the United States (five sites), Canada (three sites), Nigeria (two sites), and one site in each  Figure 3: (a) The effect of EC on the increase of CPT cone resistance and relative density (CPT data from [9]).(b) CPT tip resistant increasing with time after (1) blast densification (Mitchell and Solymar [12], Fordham et al. [13], and Jefferies et al. [14]); (2) vibrocompaction (Mitchell and Solymar [12]); and (3) dynamic compaction (Schmertmann [15]).Research and comparison have been done by Charlie et al. [10].
of the following: Japan, Sweden, and China.The soil deposits in these sites ranged from fine alluvial sand to gravel and cobblestone.The maximum ground water level was at the depth of 2 m from the ground surface and the thickness of target layers for soil improvement was between 4 and 20 m in most sites.
EC was carried out in one to four phases, at grid points forming square or triangular grid pattern.The distance between the blast holes varied, depending on the limitations of each site as regards the weight of charges.In some sites, a grid of blast holes spaced at 3 m to 6 m was used, whereas in some other sites the blast holes were at spacing of up to 15 m and loaded with higher amount of charges.Various types of charges were used in each country depending on their availability, practicality, and cost.
Awareness of the amount of energy implemented to target soil of each site is necessary to have good judgment about EC performance in every project.In order to reach the criteria that describe amount of used energy, Powder Factor (PF) can be determined as follows [16]: where PF is Powder Factor (gr/m 3 ),  is charge weight (kg), and  is soil volume that is modified by EC method (m 3 ).
It is related to thickness of improved layer (ℎ 0 ), borehole arrangement, and distances ().For example for square grid pattern,  is described by the following: This parameter has been calculated for collected cases and with a summary of the database records shown in Table 1.A brief description of each site, also, is given below.
Site Number 1: Tokachi Port Project [17].For performance evaluation of steel sheet piles to prevent lateral spreading due to liquefaction consequences of an earthquake, a complete EC test was carried out in Tokachi port, located in Hokkaido Island in Japan.The soil profile at the site was predominantly loose, fine silty sand (approx.15% fine material) of thickness of 8 m and dredged from the sea bottom.
Site Number 2: Seymour Fall Dam [8,18].In early 1960s, a new concrete wall and an earth dam with the height of 30 m were constructed on the Seymour River, approximately 18 km north of the Burrard Inlet within North Vancouver city and adjacent to the existing 9 m high concrete dam (Seymour Fall Dam).Due to the seismology of the region, EC and dynamic compaction methods were used for compaction of the new dam site between 2004 and 2005.The ground water at the dam site was nearly at the ground level, while there were sand deposits along with coarse gravel and cobblestone extending to a depth of 30 m in the zone targeted for ground improvement.After EC, an average settlement of 5% to 7% was reported in the target layer.[20].The site is located in a coastal region of South Carolina where the ground profile was classified into six layers.The target layer for EC was located at the depths of 7.5 to 13 m (5.5 m thick) from the ground level.This layer comprised fine sand with relative density between 20% and 30% and a fine content of 4%.The layer was fully saturated, being below the ground water level.EC was designed and performed at this site in 4 phases using an arrangement with square grid pattern in 8 months.The ground surface experienced an average settlement of 168 mm after first phase with a recorded settlement between 120 mm and 90 mm in phases 2 to 4. In total, the target layer for improvement settled approximately 490 mm.

Site Number 4: Foundation of Tailing Dam in Ontario
Canada-Test Area 2 [21].This project was conducted in order to improve the foundation soil for a proposed tailings dam raise in Ontario, Canada.Based on the laboratory test data, the deposited tailings were noted to be heterogeneous and they consisted of alternating layers of fine sand, silt, and silty sand.
Cone tip resistance of 0 to 5 MPa was typical of the saturated tailings at depth and relative density varied between 40 and 60 percent.The depth of target layer was about 20 meters and average of water table depth was measured as approximately 3 m below the surface.Settlements after blasting were in the range of 180 cm to 250 cm within the plan area of the test.It was reported that these settlements were concentrated in the lowest 20 m of the site, which gives an induced vertical strain of about 10%.Postcompaction penetration resistances were measured using the CPT, two months after the end of ground treatment, and they were doubled in comparison with before treatment.Figure 4 illustrates CPT results before and after EC.

Sites Number 5 and Number 6: Jebba Dam (Testing and the
Main Stages) [22,23].Compaction of the alluvial deposits at the depths of 30 to 45 m definitely brought special credit to EC. Access to the deeper point more than 30 m is usually not applicable and functional for the other common methods.If so, the work can be difficult and costly; therefore, achievement of soil improvement was recognized from 30 to 45 m by EC efficiency and it involved promising results.Jebba rock fill dam, with a height of 42 m, in Nigeria was constructed on 70 m thick alluvial deposits on the Niger River.In order to prevent both differential settlement of the dam foundations and liquefaction, the existing deposits at the depths of 30 to 45 were densified by EC.Before major stage of compaction, some EC tests were done near the main site; after that, primary place was divided into 5 zones and blast densification was carried out for each zone separately.
Based on site investigations, alluvial deposits were classified as medium-to-coarse sand mixed with gravel without any fine contents.The average uniformity index was 2.94, while the average  10 was reported as 0.31 mm.The relative density of the soil in the test zone was reported between 35% and 75%, while in the main zone it ranged from 35% to 60%. Figure 5 shows the grain size envelope of the target deposits in this project.For test site, EC was designed and performed in three phases with a checked pattern of blast holes.Blast holes were at the distances of 5 m from each other in each phase and were loaded with Noble's Special Gelatin charges   of 80%.The weight of charges in each blast hole in phases 1 to 3 was 3, 2, and 1 kg, respectively, with their center of mass planted at the depth of 36 m.The performance of EC was measured by surveying the surface settlement and comparing the CPT records before and after blasting.Final settlement of approximately 27 cm (13 cm, 9 cm, and 5 cm from phases 1, 2, and 3, resp.) was reported as a result of EC.For designing EC related to the main zone, more charges were used and the distance between blast holes in each phase increased to 10 m.
Site Number 7: Chicopee Project I [24].EC was carried out as part of the foundation design for new buildings at an industrial park in Chicopee, Massachusetts, USA.The objective was to prevent liquefaction induced by earthquakes in strata existing at depths of 6.1 to 15.2 m.The soil profile at the site comprised mainly alluvial sand layers with gravel and some silt up to the depth of 30 m.The average settlement was measured to be approximately 13 m, which was equivalent to 1.4% of the thickness of the target layer.CPT was carried out before and after blasting for evaluation of the improvement performance.As illustrated in Figure 6, upgrading in the soil strength after EC is found generally in the target deposits layer, but more increased strength around charges position at depth is certainly conspicuous.
Figure 6: CPT results before and after blast densification in Chicopee project I (CPT data: [24]).
Site Number 8: Soil Densification for a Building in Florida, Florida Job Project [25].This project was located at a site in Lakeland City, Florida, USA.Layers of fine-to-medium sand, loose and of uniform grain size distribution, existed between the ground level and a depth of 8 m.The ground water level was at a depth of 0.9 m, but the percentage of fines in the soil was unknown.The loose nature of the layer necessitated EC to improve the site prior to construction.A considerable settlement occurred in the ground surface due to consequent EC.The settlement was reported to be 0.67 m, which was approximately equivalent to 8% of the thickness of the layer targeted for improvement.This was a significant value in comparison with other projects.
Site Number 9: Franklin Fall Dam [26].Franklin Fall Dam in New Hampshire, USA, was built on the Pemigewasset River to control flooding and was completed in 1943.The dam reservoir and the surrounding areas are one of the tourist attractions in the USA.Lyman (1942) studied the performance of EC for densification of this dam and reported the method as a success.The riverbed composed mainly of fine-to-medium sand, which had been transferred from upstream locations and deposited in a loose state.Similar conditions apply to the dam site.Observations on the riverbed revealed the presence of silt and sand structure with a fine content of 25%-40%.The thickness of the target layer for improvement in this project was 6.1 m, but, unfortunately, there were no site investigation records before and after the improvement activity.
Site Number 10: Road Construction Project along Söderhamn-Enånger [27,28].The method of EC for soil densification has been implemented to a road construction project in central part of Sweden.For building this road, the natural organic soil was excavated and replaced with a fine-grained fill.This fill with varied thickness between 2.5 and 5.5 meter was subjected to compaction by blasting on three phases separated by at least 2 to 3 weeks.The fill includes about 5% clay, 20% silt, 50% sand, and 10% gravel; also there is a significant number of cobbles and boulders in till.As the ground water level was at 0.5 meter below the surface and initial dry density of deposits was 1530 kg/m 3 , the layer was saturated and loose.For this case, blast compaction has been carried out in triangular pattern with holes loaded with charges about one to three kg in weight.As a result of EC, target layer settles in range of 3% to 10% of the filling thickness.Moreover, geophysical method SASW (Spectral Analysis of Surface Waves) was used to detect changes in fill stiffness due to the blasting activities.
The results of these measurements showed some parts of the fill seemed to get firmer as well as the volume change of the deposits presents improvement in soil density.
Site Number 11: Shanghai Harbor, China [29].To determine the ability of EC to densification of the reclamation by bumping filling sand (sand, i.e., poured on seabed to settle on self-weight) foundation, a series of in situ trials were carried out in a harbor in Shanghai.All field tests were carried out in a port which was formed by bumping fine clean sand with coefficient of uniformity about 2; thicknesses of sand layer was 10 m and mean ground water level of trial field was −0.7 m-0.9 m, so this layer is loose and saturated.In two of these trial tests (T7 and T8), EC was designed in two separate types of coverage (second type of coverage was carried out 7 days after the first) with square plan as shown in Figure 7.
A record of monitoring T7 test has been reported completely indicating about 10 cm settlements during 28 days which was observed at center of holes plan and cone resistance approximately doubled along the target soil due to EC.

Site Number 12:
Quebec HQ SM-3 Dam, Canada [30].A large EC project was carried out at SM-3 site along the Sainte Marguerite River, Quebec, in 1995.In this project, a 100 m by 120 m area with depth of up to 20 m of riverbed was densified in order to reduce the potential for static liquefaction and improve the stability of an excavation for cofferdam during construction of main dam.Site investigation showed that the sand deposits at the site consist of loose sand overlying a dense sand layer with a fine content less than 5%.The EC program improves the relative density throughout the site to an average of 75%.In one area of site, CPT was performed before and after blasting at intervals of 2, 12, and 35 days, which is shown in Figure 8.

Site Number 13:
Test Site in Oakridge Landfill, South Carolina [31].The test site was located at the Oakridge Landfill in Dorchester County, South Carolina (approximately 50 km northwest of Charleston).Blasting was used to densify a layer of potentially liquefiable loose sand along the perimeter of the landfill to address slope stability concerns during a seismic event.Target layer is a loose layer of fine sand known as the Raysor formation with approximate 4 m of thickness.The initial relative density of the loose black sand was estimated to be 15-30%.Four blast events took place over the course of 21 days (7 days between blasts).The blast pattern and locations of instrumentation for the test section are shown in Figure 9. Blast holes were spaced 6.1 m apart and contained 15.5 kg of Hydromite buried at a depth of 10.7 m.Finally, about 0.3-0.6 m settlement occurred in the target zone after four explosion stages.Relative densities near the center of the zone appear to have increased to values between 80 and 90%.Despite significant amounts of densification, penetration resistance measured with the CPTu indicated no improvement.

Geotechnical Aspects of EC
A useful way to estimate the results of EC is to analyze case histories of sites improved by this technique.In the following paragraphs, an opportunity has been taken to analyze a database of case histories in three categories including (1) zoning deposits for EC, (2) initial stability status, and (3) EC performance on soil penetration resistance; this is simply because capability to predict and analyze changes in the ground properties is of interest to geotechnical engineers and researchers.
4.1.Zoning Deposits for EC.Engineers may be faced with the challenge of identifying suitable soil types for EC.An ability to do this will help reduce the number of preliminary tests, necessary to be carried out, thereby, leading to reduced project costs and time savings in construction.Considering the available data from compiled sites, the performance of EC and the efficiency of this method are evaluated under various geotechnical conditions in terms of material and grain size distribution using the database records.This evaluation according to successful performance of EC can give a good insight into the performance range of EC realizing to the type and fabric of deposits.Figure 10 shows grain size envelope of some sites in the database.Undoubtedly, problematic soils are affected significantly by improvement measures.This is true not only for EC but also for other soil improvement methods.Various factors such as the type and grain size distribution, initial density, and saturation ratio of the soil are important for the result of EC.So far, a specific range of grain size has not been given for the performance range of EC.Some researchers [7,32] have suggested that the range of soil types suitable for vibrocompaction is also suitable for EC.According to this suggestion, saturated sands with 20% silt content and less than 5% clay are appropriate for EC.Other researchers, for example, Bell [33], has reported that soil having up to 70% silt content or 10% clay is suitable for EC.However, so far no significant upper limit of fine content has been recommended for EC.Plotting the grain size envelopes of the sites in the database in a single graph, as shown in Figure 11, may give broad indication of the performance range of EC.To obtain this boundary, it is assumed that grain size and distribution do not change after EC; however, it is realistic to expect a few changes in grading because crushing, initial condition of deposits, and soil type have more effect on EC consequence that is remarkable.

Initial Stability of Target Deposits. Cone penetration test
(CPT) is one of the in situ tests done in a site investigation activity.CPT is fast and cost effective and it provides continuous measurement of soil properties with depths.CPT data are used for investigation of deposits initial condition before.In this paper, two approaches are used for evaluation of initial stability status of deposits as follows.

𝑄 tn
Criteria.The dilatancy behavior of sands is affected by mineralogical characteristics and grain size in addition to placement density and confining pressure.It is expected that the factors affecting dilatancy behavior also affect measured cone bearing.However, it is not clear whether they are affected in the same manner.Sladen and Hewitt [34] defined  tn as a border between dilation and compression behavior  based on CPT results, which is determined according to the following: where  tn is normalized total cone bearing stress,  t is total cone bearing stress after pore pressure correction (for clean sands and gravels  t =  c ), and   V is vertical effective stress.Sladen and Hewitt's [34] criterion for the dilativecontractive boundary is  tn = 70 bar (1 bar = 0.1 Mpa) and sands with less than 70 bar are considered loose or contractive while sands with greater than 70 bar are considered dense or dilative.Pincus et al. [35] by studying the results of resistivity cone penetration testing (RCPTu) in various sites proposed  tn = 55 bar for loose-dense boundary.Therefore, they suggested that if  tn < 55 bar then the sand is loose and susceptible to liquefaction, and if  tn > 55 bar then sand is dense and its liquefaction susceptibility is very low. tn values in treated layers before blasting for studied cases as well as reference lines according to Pincus et al. 's and Sladen and Hewitt's criteria are presented at Figure 12.

Soil Behavior Classification Charts and Liquefaction
Zone.One of the main applications of CPT is soil classification and some researchers have presented graphs for using CPT records in the identification of soil types and intrinsic conditions.Eslami and Fellenius [36] catalogue several existing CPT-based soil classification charts and compared their relative reliability.The graphs are based on normalized plots of cone resistance  c , cone sleeve friction  s , and pore pressure  2 , and they could be used for identification of problematic soils suitable for improvement by EC.For evaluation of the conditions of target deposits before blasting by using the soil classification graphs, sites numbers 3, 6, and 7 in the database were selected.In these sites, the performance of EC includes promising outcomes via cone penetration resistance and relative density of target deposits increased significantly after EC.In this study, charts proposed by Robertson et al. [37] and Eslami and Fellenius (2004) have been utilized for CPT-based soil classification.Hence, location of CPT records of selected sites in the classification graphs, which is shown in Figure 13, can be implemented as a criterion for efficient implementation of EC on specific geomaterial deposits boundaries.

EC Performance on Soil Penetration
Resistance.Generally, CPT can be used for evaluation of EC before and after blasting which is shown in Figure 14.It is also common to perform CPT tests in different time intervals after explosion because penetration resistance increases by time so, in this study, the last reported CPT data has been utilized for each site.A combination of CPT data and experimental equations may be used to determine other geotechnical parameters such as relative density (  ) which is calculated by (4). Figure 15 shows a summary of CPT records of compiled sites before and after soil improvement (see [38]).
where  c1 is normalized cone resistance,  OCR is adjustment factor for overconsolidation ≈ 1, and  Age is adjustment factor for age ≈ 1.  Shanghai Harbor Jebba dam South Carolina-Test site q c (MPa) q c (MPa) q c Figure 14: CPT records before and after EC in some sites in the database.

Comparison and Discussion
As explained in brief for compiled cases, the efficiency and successful issues related to geotechnical aspects have been proved for EC performance including increasing resistance (strength), reducing volume change (settlement), and upgrading internal stability.Therefore, most of these criteria have been adopted for the database records.
For all cases, the soil layers targeted for improvement were 100% saturated or so were mostly made up of sand with a fine content of almost 0%-40% and in a loose state which means   was almost close to 50% or less as it is shown in Figure 15.However, some of the target layers had gravel and cobblestone or rubble with diameters of more than 1 m (e.g., in the case of Seymour Fall Dam project).Successful improvement indicates the capability of EC in a wide range of soil deposits.The proposed range for soil improvement by EC in this study may be used as a guide for selection of the EC option among other deep improvement options.
Accordingly, suitable performance of EC is expected from gravel to silty sand with a silt content of less than approximately 40% and clay content of less than 10%.
However, the best EC, which means significant increase in relative density, penetration resistance, and especially diminution of liquefaction potential, was achieved with fineto-medium sands with a fine content of less than 5%.Other determining factors in the selection of soil improvement method are (i) environmental conditions, (ii) depth of the In some sites, the treated layer experienced significant volume reduction, but penetration resistance did not increase significantly.Cases 3 and 13 are good examples for this phenomenon.After blast densification at a test site in South Carolina, USA, large surface settlement of about 8% of the layer thickness happened, while average values of  c increased only to about 28% after 1034 days.Likewise, in Oakridge Landfill vast volume changes and increase in   can be seen whereas CPT records do not show significant changes because of EC.It is concluded that some reasons such as confining layer above target soils can affect CPT results because this layer is an obstacle for escaping gases generated by blasting.
Results of investigation show that EC was carried out in loose deposits with  tn < 55 because these types of soils are susceptible to liquefaction.Hence,  tn criteria can be used to identify deposits, which are suitable to be improved by EC.
In some sites such as Chicopee I, Quebec and Shanghai project, the fine content was less than 5% and this enabled achievement of a significant increase in the cone resistance.
In general, according to data of Table 1 and comparisons presented at Figure 1, it can be concluded that the cone resistance will increase up to twice of initial resistance value after blast densification in hydraulically deposited alluvial sands having a fine content of less than 5%.
As for the soil classification graphs based on CPT records, which are shown in Figure 13, EC had promising results for soils falling in the silty clay to silty sand zones defined by the Robertson et al. (1986) chart and also in the silty sand zone defined by Eslami and Fellenius (2004) chart.Therefore, it can be concluded that these zones in classification charts represent soil types that have liquefaction potential.Through comparison of these two charts, it may be seen that a more suitable scattering of records is achieved by Eslami and Fellenius (2004) chart as the records fit well in the sand and silty sand zone.Nevertheless, in the chart of Robertson et al. (1986) the data records are spread over more zones, making it difficult to interpret the range of grain size distribution for soils suitable for EC treatment.

Summary and Conclusions
EC for ground modification covers a wide range of soil types with regard to the grain size distribution.In this study, thirteen successful case histories in USA, Canada, Nigeria, Japan, Sweden, and China have been collected and studied, focused on the range of grain size distribution.The density of target deposits for improvement has been evaluated with regard to the settlement of layer and increase of strength, which were obtained from the cone resistance, relative density, and visual observations reported in the sources.Consequently, these geotechnical variations indicate the successful performance of EC for improving the soil deposits.The geomaterials of sites mainly comprised fine-to-medium sand hydraulically deposited alluvial layers, silty sand, or sands mixed with gravel and cobblestone, while the existing fine content grains, which are less than 0.075 mm in size, in the site's deposits were different between 0 and 40%.
Design of EC in these sites was carried out in square or triangular grid patterns and in one to four phases.The distance between the blast holes varied depending on constraints of each site in terms of application of explosives.Study of the grain size distribution of target deposits for improvement and design status of EC, in addition to analysis of CPT records before and after improvement have been used and the following results are inferred for zoning: (i) EC was performed successfully in a wide range of soil types, from coarse gravel with cobble or rubble with a diameter of 1 m, to silty sands with 40% silt content and 10% clay.The best performance of EC was in alluvial sands with a fine content of less than 5% and hydraulically deposited layers with an initial relative density of 30% to 60%.Performance range of EC is in all soil types with potential liquefaction, which covers a wider range of soil types in comparison with other soil improvement methods like the vibration methods.(ii) Cone penetration test (CPT) can be used as an evaluation tool for EC along with the measurements of settlement.Analyses have shown that, through EC, the cone resistance of sand deposits easily increases up to 200% with a fine content of less than 5%, which would be followed by 10% to 30% increase in the relative density of soil.The increase of the fine content and density of deposits would affect the performance of EC in relation to the variation of the penetration resistance of the soil.(iii) Deposits with CPT records fitting into zones 6 to 9 of Robertson et al., 1986, classification chart and zone 4 of Eslami and Fellenius, 2004, chart are the best for improvement using the EC method.Less scattering of CPT records was observed in Eslami and Fellenius's chart, which gives more assurance with identification of target deposits for EC.(iv) In addition to the position of problematic deposits with regard to grain size distribution and initial strength, other important parameters like depth of target layer, relative costs, and environmental limitations may play an important role in selection of EC as an alternative for deep improvement among other methods.

Figure 1 :
Figure 1: Wave propagation caused by explosion shock, soil liquefaction, and settlement.
N weeks/tip resistance of one week (b)

Figure 4 :
Figure 4: CPT results before and after EC in test site [21].

Figure 5 :
Figure5: Grain size envelope of the existing deposits in the testing area and site of Jebba dam project[22].

Figure 8 :
Figure 8: Time dependency of CPT cone tip resistance  c at the Quebec HQ SM-3 Dam [30].

Figure 9 :Figure 10 :
Figure 9: (a) Pattern of blast holes and location of instrumentation.(b) Profile of target layer [31].

Figure 11 :
Figure 11: Suggested grain size zoning for soil improvement by EC.

Figure 12 :
Figure 12: Initial status of target deposits based on  tn criteria.

Table 1 :
Database records summary.
Figure 15: CPT and   Results before and after EC in some sites in the database.targetlayerforimprovement, and (iii) grain size distribution of the deposits.For example, for the great depth of the target layer encountered in the Jebba dam project, EC is the only alternative that is appropriate.Considering Figure15, it can be noticed that EC has increased the average cone resistance by 30%-200% in most sites.Besides, the relative density in all deposits increased by 10% to 50% due to EC.The changes in the cone resistance values vary from one site to another because factors such as the CPT recording time, applied energy, soil type, and initial strength also influence the final penetration resistance after blast densification.