A new sand raining technique to reconstitute large sand specimens

1 Introduction

The methods used to prepare samples have an effect on the soil grains. Studies found that the application of dynamic compaction to sand soil causes crushing to its grains. Because of its capability to produce wide ranges of sand relative densities, the raining sand method has gained popularity [1,2,3,4,5,6,7,8]. Furthermore, when large soil samples must be reconstituted fast for calibration chamber tests or sample static or dynamic tests for different geotechnical laboratory models (e.g., shallow and deep foundations and retaining structures), raining method proves useful [9]. Kuerbis and Vaid [10] reported the main advantages of raining method. They stated that in the raining method, the sand grains are not broken as they are in traditional techniques. They also mentioned that in this technique, the ease with which a large number of samples can be prepared; moreover, the ease with which preparation can be completed.

In the raining technique, the required relative densities of the soil samples are controlled by factors like the diameter of the rain hose (through which the soil travels) and the height of rain sand. By changing these factors, different relative densities can be attended [11,12,13,14]. The researchers prefer the free-falling in the air raining method. This method involves raining sand in a dry condition exposed to the air during the falling. By such a process, samples of various sizes can be implemented for laboratory examination. This can be conducted by controlling the falling height and pouring dry sand particles through the air in the sample container at a constant speed (assuming the size of the sand-falling aperture remains constant). The relative density (RD) in this method depends on factors like the deposition intensity (DI) (which is affected by the rain opening diameter), the falling height (HF), the gradient and properties of the soil grains, etc. [9,11,13,15,16].

The meaning of the HF is the distance between the bottom of the rain opening diameter to the surface of the sand in the sample [3,4]. According to the studies of Rad and Tumay [11] and Vaid and Negussey [17,18], the HF has a significant effect on the values of RD of the soil grains. These researchers found that there is a non-linear relationship between HF and RD, where the increase in fall height leads to an increase in sand particles' collision velocity. This increase continues until a critical height is reached, at this critical height, the height of rain becomes ineffective in increasing the RD of the soil samples [19]. In addition, there is an almost linear increase in critical velocity when the sand grain size is increased [15]. The HF has a major in controlling the RD of the sample while the influence of DI is negligible. The increase in DI increases the porosity of the deposition [15,16], which leads to an increase in the void ratio [11], that is, the volume of the fallen sand is larger, which increases the chance of sand grains colliding before they fall into the container, thus lowering the RD value in the soil sample [9,12,20]. Therefore, using a soil rain nozzle with smaller apertures results in higher RD values (in the case of free-falling soils) [11].

Previous studies highlighted the fact that the raining methods suffer from some disadvantages. In such methods, it is difficult to obtain RD larger than 70–75% [12,14,21,22]. According to literature [12,23,24,25,26,27,28], experiments with different heights of the sand drop, HF, have been conducted to show the change in the RD. The finding of these experiments has been investigated in this article; based on these finding, the relationship between RD and HF has been plotted as shown in Figure 1. It has been noted that as HF increases, the RD rises as well, owing to the increased soil density. This growth is apparent until the critical height (about 75 cm) is reached, at which then the height becomes ineffective. Reexamination of Figure 1 shows that the maximum RD values abstained from sand raining are in the range of 70 and 75%.

Figure 1

Variation of RD and HF in raining sand method from different literature.

Researchers have been tried to improve the efficiency of the sand raining method as an attempt to achieve higher RD. Different densities were obtained either by controlling the HF or DI. However, there are different approaches that have been adopted by the previous researchers to overcome the difficulty to obtain high RD, [9,13,14,29]. Some researchers used a set of diffuser sieves in the sand raining devices to obtain a very dense specimen. Other researchers showed that wide ranges of densities can be obtained using a portable curtain rainer to sand specimens. Also, it was proposed that the combination of the height of rain and the DI may provide a way to achieve higher densities. Also, the use of a rainer curtain and diffusing sieves in the raining device was found to be effective to achieve high densities.

The raining technique requires to be more investigated for its use in successfully preparing reconstituted large experimental sand models with relatively very high relative densities at a lesser time. The present work marks out details of a new raining system where both the height of rain, HF, and the DI are simultaneously controlled. The new system is a trapped air raining device (TARD) competent for reconstituting soil samples with different ranges of relative densities, from loose to very dense. It consists of a cylindrical sand tank, a top hose (with metal funnel) to fill the tank with the sand, a lateral hose to enter the trapped air (from an electric air compressor), three own valve rain nozzle with different opening diameters at the bottom of the tank to rain the sand, and pressure gauge. Under varied trapping air pressures (0, 1, 2, 3, 4, and 5 bar), the HF (30, 60, and 90 cm), and employing rain nozzles with three different opening sizes (0.5, 1.0, and 2.0 cm), the new system was used to investigate the effect of trapping air pressures on the RD and DI for SP-SM soil rained from different HF.

2 Description of new raining device

The new raining device named “TARD,” consists of a 77L capacity cylindrical sand tank (made of steel with a thickness of 1 mm) with five nozzles, each with its own valve. The top one with a hose connects with a metal funnel to enter the sand to fill the tank with sand. The side or lateral nozzle connects with a special rubber tube designed for high air pressure to enter the trapped air into the sand tank. The rest nozzles are three own valve rain nozzles with different opening diameters (0.5, 1.0, and 2.0 cm) at the bottom of the tank to rain the sand. A pressure gauge attaches to the top lateral side of the sand tank to measure the pressure of the trapped air. To provide the sand tank with confined air, an electric air compressor is used, as shown in Figure 2. The TARD can be mounted to a ceiling using three metal suspension chains connected to the top of it. With TARD, sand can be rained with the assistance of a high air thrust (more than 6 bars) with the possibility of changing the HF and controlling the diameter of the raining nozzles. With TARD, there is no ability to clog the sand raining nozzle, as in the traditional raining sand method, because the pneumatic thrust prevents sand grains from agglomerating, as well as the ability to use larger diameters for the nozzles. With TARD, there is an ability to fill large and multiple samples in a short time due to the presence of multiple nozzles to drop sand, as well as the high compressed air thrust that accelerates the flow of sand.

Figure 2

TARD details.

3 Experimental work

3.1 Properties of soil for samples preparation

The soil utilized in this investigation to prepare the samples was a dry, rather silica sand with grain diameters ranging from 0.075 to 2 mm. Figure 3 shows the grain size distribution curve of tested sand. The investigated sand particles have a relatively low specific gravity of 2.45. According to ASTM D2487, the Unified Soil Classification System, the investigated sand is classified as poorly graded sand with silt (SP-SM). The sand tested has a uniformity coefficient and a curvature coefficient of 4.12 and 1.13, respectively. Also, it has an effective grain size (D10) and a mean grain size (D50) of 0.10 and 0.33 mm, respectively (ASTM D422). The maximum and minimum dry densities of sand, for estimating the RD, are 1.78 and 1.41 g/cm³, respectively (ASTM D4253 and D4254).

Figure 3

Grain size distribution of sand.

3.2 Raining test procedure

The experimental program of this work consists of 54 total raining tests carried out in two main parts. In the first part, nine raining tests were conducted using TARD without applying air pressure (no trapped air). In these tests, rain nozzles with three different opening sizes (0.5, 1.0, and 2.0 cm) were used to reconstitute sand soil samples. With each of these opening sizes, three HF were used (30, 60, and 90 cm), and these heights were kept fastened in each test. The values of RD for each trial were determined using small steel containers of 375 cm³ volume. These containers were placed inside a steel box of 70 cm in length, 70 cm in width, and 60 cm in height to evaluate the purposes of this work as shown in Figure 4.

Figure 4

Details of raining test with TARD.

The second part of this work includes 45 raining tests using TARD. To reconstitute the soil samples in this part, five varied trapping air pressures, TAP, of 1, 2, 3, 4, and 5 bars were applied. The heights of raining are as shown in the first part, 30, 60, and 90 cm. Also, for each raining height, three different opening sizes (0.5, 1.0, and 2.0 cm) were employed to control the DI. The new system was used to investigate the effect of trapping air pressures on the RD and DI for SP-SM soil rained from different HF. Moreover, to obtain the RD, the same steel containers were placed inside the steel box shown in Figure 4.

4 Results of testing and discussion

In this work, to obtain the high values of RD of sand samples for laboratory reconstituting of the physical model, comprehensive laboratory investigation was conducted using a new raining device named “TARD.” This work includes analyzing the unit weight of the dry sand to calculate the RD in two different parts, discussed in the previous section. These parts include the raining of sand using TARD without trapped air (zero air pressure), while in the second part, the different trapped air pressures were applied during the sand raining process. For both parts, under different pressures, the new system was used to investigate the effect of the height of rain, and rain nozzles opening sizes on the RD values. The result of this work was presented and discussed in this section.

To investigate the efficiency of TARD on RD preparation, first, pluviation tests were carried out controlling the HF and DI under zero TAP, as shown in Figures 5–10. For SP-SM, soil raining through a nozzle of 1.0 cm diameter opening produced an insignificant effect on RD values, even though HF increases from 30 to 90 cm (Figures 5 and 7). A more pronounced effect was noticed using nozzle opening diameters of 0.5 and 2.0 cm as shown in Figure 6. Figures 8–10 present the DI values for sand raining through different nozzle diameters opening. For sand rain from HR of 30 cm through a nozzle diameter opening of 0.5, 1.0, and 2.0 cm, the DI values are 0.028, 0.087, and 0.095 g/cm³/s, respectively. For the same range of nozzle diameter opening, an insignificant effect of HF increasing on DI values can be noted where DI ranged from 0.022 to 0.078 g/cm³/s.

Figure 5

Variation of RD, TAP, and HF for nozzle size of 0.5 cm.

Figure 6

Variation of RD, TAP, and HF for nozzle size of 1.0 cm.

Figure 7

Variation of RD, TAP, and HF for nozzle size of 2.0 cm.

Figure 8

Variation of DI, TAP, and HF for nozzle size of 0.5 cm.

Figure 9

Variation of DI, TAP, and HF for nozzle size of 1.0 cm.

Figure 10

Variation of DI, TAP, and HF for nozzle size of 2.0 cm.

The effect of variation of RD for SP-SM samples rain under zero TAP can be seen in Figures 5–7. It is clear that with nozzle opening diameters of 0.5 and 2.0 cm and at zero trapped air pressure, the effect of the height of rain on the RD of the sand was more pronounced than that of the nozzle opening diameter of 1.0 cm. However, a low RD value of <40% was achieved for RD of 30 cm and nozzle opening diameters equal to 0.5 and 2.0 cm. Also, the maximum RD (62%) was obtained using HF 90 cm and a nozzle diameter of 0.5 cm. Further increase in nozzle diameter produces low RD values, from 34% to 61%.

As can be seen, very high RD cannot be achieved by controlling the HF and DI under zero TAP. Thus, further experiments were conducted in an attempt to achieve the required RD values, different TAP were applied in these experiments. As shown in Figures 5–7, considering TAP from 1 to 5 bars, varying the height of rain from 30 to 90 cm, and nozzle size from 0.5 to 2.0 cm resulted in different changes in RD values. The effect of TAP on the RD values is very clear. There is non-linear relation between TAP and RD for different HF and nozzle sizes. The second-order polynomial showed the best fitting to this relationship with very high R ² values of range from 0.933 to 0.999.

As shown in Figures 5–7, for each HF value, there is an increase in RD with the increase in TAP; however, in the range of 30 and 60 cm, the HF has a minor in controlling the sample's RD. However, under different TAP values, HF 90 cm shows a very pronounced effect on RD value for nozzle size less than 2 cm. On the other hand, the effect of the height of sand rain on RD values decreases with increasing the nozzle rain opening diameter and the pressure of trapped air. This effect has vanished for sand rained through a nozzle opening of 2 cm diameter and under trapped air pressure of 5 bars. The potential of reaching a very high RD (99.7%), i.e., achieving the maximum dry density, was attained using a pressure of 4 bars and a sand rain height of 90 cm, by raining the soil from a nozzle with a diameter of 1.0 cm. A comparison with previous studies (shown in Figure 1) shows that when reaching the critical height of HF, the increase in the RD is very slight, no matter how high the height is, and therefore the difficulty of obtaining the RD exceeds 75%.

The air pressure provides high momentum to the soil grains, as a result, the colliding of grains between each other is reduced., thus, to a certain extent, the RD increase. This is depending on the pressure of trapped air and the raining height of sand. While in the classic raining method (zero TAP), when the height of rain increases, the falling of sand grains resists by air during the falling, in addition. Also, the bumping of sand grains into each other causes a reduction in both the speed and momentum of grains, and as a result, the RD of sand reduces. Figures 8–10 present the DI values, DI, for sand raining through different nozzle diameters opening under different pressures, TAP. As shown, regardless of HF value and nozzle size, there is a direct non-linear relationship between the TAP and DI. It was found that raining the SP-SM soil rain from a drop height of 30 cm, under 5 bar pressure, and through a nozzle diameter opening of 0.5, 1.0, and 2.0 cm achieve the maximum DI values (0.840, 1.706, and 1.682 g/cm³/s). Compared to the classic raining technique (zero pressure), this represents an increase in the value of the DI ranging from 19 to 30 times. For the condition of maximum RD value (99.7%) that was achieved for soil rain from 90 cm height, under 4 bar TAP, and through 1 cm nozzle size, the obtained DI value is 0.89 g/cm³/s, which indicated a reduction in preparation time of more than 90% in comparison to classic raining technique. Authors [9,11] indicated that the deposition porosity and the void ratio increase when DI increases. As a result of such increasing, the volume of rained sand increases, and therefore, the RD value decreases. Based on what was achieved through this study, such a negative effect of DI increasing was overcome using the new device “TARD” by applying trapping air pressures to the soil during the process of rain.

5 Conclusions

The TARD, designed and utilized in this work, composed of rain nozzles with different opening diameters for sand raining and trapped air to accelerate the flow of sand. In the present article, TARD was used to rain poorly graded sand with silt (SP-SM); the following are the main principal conclusions:

1. “TARD” is used to reconstitute sand samples for different geotechnical laboratory models. To prepare sand samples for a specified RD, an appropriate combination of the height of rain, nozzles opening diameter, and pressure of trapped air might be chosen.

2. Wide ranges of relative densities of poorly graded sand with silt, from loose state to very dense state, may be produced using TARD with different combinations of the height of rain, nozzles opening diameter, and the pressure of trapped air. The loose state of the sand was achieved using a height of rain equal to 30 cm, the nozzle opening diameters equal to 0.5 and 2.0 cm, and at zero trapped air pressure. While very dense states were reached when the soil rain from a height of 90 cm, through 1.0 cm opening diameter, and under trapped air pressure of more than 3 bar.

3. With nozzle opening diameters of 0.5 and 2.0 cm and at zero trapped air pressure, the effect of the height of rain on the RD of the sand was more pronounced than that of the nozzle opening diameter of 1.0 cm. Also, the influence of the height of rain on the RD of sand is insignificant for nozzle opening diameters less than 1.0 cm and trapped air pressure less than 3 bar.

4. At a higher height of sand rain (more than 60 cm) and a nozzle rain opening diameter of more than 1.0 cm, less effect of trapped air pressure on the RD of the sand was more pronounced.

5. The effect of the height of sand rain on RD values is decreased with increasing the nozzle rain opening diameter and the pressure of trapped air. This effect has vanished for sand rained through a nozzle opening of 2 cm diameter and under trapped air pressure of 5 bar.

6. TARD was found appropriate for reconstituting SP-SM with very dense sand samples achieving higher DI values and a reduction in preparation time of more than 90% in comparison to the classic rain technique, thus it is time-saving.