Scale Tests to Estimate Penetration Force and Stress State of the Silica Sand in Windfarm Foundations

: The analysis of the soil behavior when the pile is driving into the seabed in offshore wind platforms is one of the major problems associated with this new form of clean energy generation. At present, there are no scaled studies carried out analyzing the mechanical and deformational behavior of both the material of the pile supporting the engine (large steel hollow piles with a di ‐ ameter of 8 m and a thickness of 15–20 cm) and the soil where the pile is driven. Usually, these elements are installed on sands with a very small grain size displaced from the limits of dry–wet beach (water limit) toward the offshore limits, which prevents them from returning to their pre ‐ vious location in a natural way. This paper presents results obtained from scale tests in a steel pool to analyze the behavior of the sand where the piles were installed. First, the California Bearing Ratio (CBR) test was carried out to estimate the soil behavior in similar conditions to the steel pool. The scale tests consisted of the penetration of the steel tube into the sand using a hydraulic press. The objective was to compare the results for three tubes with different diameters, three different speeds, and two kinds of ending on the extreme of the tested


Introduction
The target of this article was to analyze the mechanical and deformational behavior of the soil and study the results from scale tests carried out to estimate penetration force into and stress state of the soil, as well as its application to real driven works for monopile foundation. At present, the monopile foundation covers 81% of the foundation typology of all wind farms installed in the world (Figure 1). One of the important advantages of this typology is its cost of fabrication, which is cheaper than other typologies such as jackets. In addition, the installation of a monopile is easy to execute [1,2]. However, there are no full-scale trials at sea analyzing the behavior of the seabed, and the trials carried out onshore did not take into account the geotechnical characteristics of the soil being very different to that existing in the seabed. This is one of the conditioning factors, along with the several phenomena (such as marine currents, high pressures due to great depth, corrosion, and etc. [3,4]) that lead to monopile foundations being currently oversized to ensure the stability of the structure. These structures reach weights of about 600 tons, diameters of 8 m, and thicknesses up to 15 cm [5][6][7][8][9][10][11][12][13]. In many cases, these structures are driven up to 30 m into the seabed [14]. It is worth determining the behavior of the base material on which the turbines are installed (sands). These sands are transported from the coast and settle on the seabed over time [15][16][17][18][19]. The sediment is composed of different grain size fractions, and each fraction has a particular distribution across the littoral profile. The maximum in the relative distribution is located around the shoreline (0 m depth), which decreases toward the dune and offshore, up to 6 m depth, where its percentage increases seaward again. The coarsest fraction (>350 μm) shows a discontinuity in its distribution because it is hardly represented in the sediment between 6 and 10 m depth. The finest fraction (100-l50 pm) shows a continuously increasing distribution from the shoreline in an offshore direction, and the gradient becomes steeper at 600 m from the shoreline [20].
Sand with a grain size >500 mm was not supplied by the factories. Therefore, a size of 760 mm was chosen. In addition, the D50 was negligible compared to the tube's thickness.
At present, a large number of tests such as those of PISA (Pile Soil Analysis) [15] have been carried out. The PISA project involves entities such as the Royal Geographical Society, British Petroleum, or the Norwegian Institute. The PISA project consists of several tests of scale pile driving to determine the behavior of the soil. These tests, for a foundation study of marine turbines, were carried out onshore ( Figure 2).  Figure 3 shows an overview of activity for the PISA project [15]. The PISA project features important activities such as the design of field tests, field tests, validation, design method (FE), or full-scale prediction of onshore tests. In this paper, the red area in Figure  3 was the focus for making changes (described in Section 2.2) with respect to the PISA project (site investigation, design of field tests, draft parameterized method, calibration, field tests, validation). The analysis of the results of this article provide the opportunity to develop a future project. Some finite element models were carried out with Plaxis (special software with finite elements for soil mechanical behavior). Using this software, we could analyze scale tests inside a pool. It is possible to model several sizes and diameters of the sand particles, as well as the drive simulation of piles, in dry and submerged conditions. In addition, submerged tests inside the laboratory were made in the presence of water.
One of the most important conclusions obtained from the PISA project was that the application of a design approach indicated potential for a significant reduction in design conservatism and substantial savings for selected design scenarios. In addition, the field tests and the supporting site characterization delivered a new industry standard database against which design models in clay and sand may be compared, developed, and validated. Lastly, the adoption of the PISA design approach is likely to result in reduced conservatism in monopile design and better economies for wind farm development.
However, the existing research findings on this topic are not sufficient due to the difficulty in producing a scale test. Thus, the main objective of this paper was to produce some scale tests which include characteristics and restrictions identified by PISA, as described in Section 2.2.1.

Previous Tests
The sand chosen was silica sand due to its similarity to sea sand. Four tons of silica sand and a ton of limestone sand were used because silica sand does not have a large amount of fines, whereas sea sand contains fines [20]. Each test included three samples to improve the accuracy of the results, and all of them were tested using the same parameters of the scale tests.
The sand's parameters were the same as those used in the scale test in the laboratory. There are no optimal soil conditions for these tests. The moisture content of the samples was not defined because it cannot be chosen during in situ sea tests.
In this paper, we carried out a granulometric analysis, density test, direct shear test using a direct shear test apparatus, triaxial shear test using the cell of a triaxial apparatus, and tensile test.

Granulometric Analysis
The study was carried out according to the UNE-EN 933-1 standard [21]. The objective was to determine the basic granulometric characteristics characterizing the sand, such as the granulometric curves ( Figure 4). The vertical axis shows the percentage of grains passing through a mesh according to UNE 103101:1995 and ASTM D6913 [22,23]. The most important parameters obtained were as follows:  D50 is an important parameter for the characterization of sand and the definition of coastal movement [24,25]. Furthermore, this parameter allows comparing the sand with other tests carried out at sea. Well-graded soils have coefficient of curvature values between 1 and 3, and a sand with a uniformity coefficient less than 3 is considered very uniform [22].

Density Test
The density test was carried out according to the UNE-EN 1097-6 standard [26] for sand samples. Three samples of sand were tested. Table 1 shows the values of the density of the mixed sand.   [28].
For the preparation of the sample, the sand was poured into a box, ensuring a great enough fall height to prevent compaction ( Figure 5). This option was chosen to test the material in conditions that would be found on the seabed, on which the material is deposited by its own weight. According to the standard, consolidation occurs almost instantaneously in sandy soils; hence, a speed of 1 mm/min was chosen with a break lasting between 5 and 10 min.  Figure 6 shows the graph of the tension relationship, which determines the friction angle and the cohesion ( Table 2).   [29] for silica sand samples (Figure 7). According to this standard, consolidation occurs almost instantaneously in sandy soils; hence, the time of consolidation was 1-3 h. The Young's modulus (E) of the sand was calculated as approximately 21,570 kN/m 2 . This can be obtained as a function of the triaxial test values, either at the beginning of the stress-strain curve or at 50% of the maximum stress. In this case, it was estimated at 50% of the maximum stress, as typically employed in this sort of test [30][31][32].

Tensile Test
The tensile test was carried out according to the UNE-EN ISO 6892-1:2020 standard [33] for the material of the tubes tested (Table 3). In the PISA project [15], 28 piles were tested with varied diameter, length, and wall thickness using both monotonic and cyclic loads. The field tests represent a new industry standard database against which design models for piles in clay and sand may be compared and validated.
Ideally, full-scale offshore tests using 6-10 m diameter piles would be used to benchmark the new design methodologies. However, due to the high cost of testing offshore and the technical constraints of equipment that can be mobilized, onshore testing was accepted at reduced scale. Scaled pile geometries and loading regimes were adopted, which were representative of offshore wind foundations (Figure 8). where L is the embedded length, D is the diameter, t is the thickness, h is the non-embedded length, and z is the depth reference.

CBR Test
The Californian Bearing Ratio Test (CBR Test) was carried out according to the UNE 103 502 standard [34] for silica sand samples [28].
The objective of this test was to determine an estimation of sand penetration using a piston with similar characteristics to the piles of the scale test. The CBR test was carried out using three types of compaction (15,30, and 45 hits), as outlined in the standard (Figure 9). These obtained results can serve as an estimate for the raft tests considering relevant aspects from the scale results [35].  These pool dimensions were enough to avoid the contour effects with the side walls and the floor Furthermore, there is no standardized test to be found in the ABS or DNV-GL.
The gauge layout is shown in Figure 11. The gauges were placed on strips perpendicular to each other, installed at a height of 50 cm halfway between the center of the pool and the side, as can be seen in Figure 11.
Three tubes were employed with the characteristics described in Table 4.

G1
As can be seen in Table 4, the thicknesses chosen for this test achieved the restrictions given by the PISA test [36] (mentioned in Section 2.2.1) The tubes chosen were not excessively large, and their thicknesses can easily be found in factories. In addition, the tubes were closed at one end. A hollow position denotes where the driving is carried out using the open side, and the flat position denotes where it is carried out using the closed side.
The penetration speeds for the tests were different from the PISA project, where D/300 per minute and D/500 per minute were chosen (D = diameter). In our case, these speeds reproduced very long tests. Therefore, an estimate was made on the basis of the three diameters, and speeds of 5, 10 and 20 mm/min were obtained with test durations of 1 h, 30 min, and 15 min, respectively [37]. The depth of penetration chosen was 325 mm. This length was sufficient to determine the mechanical behavior of the steel tube with the sand and, thus, avoid contour effects.

Scale Test
This test consisted of the penetration of the steel tube into the sand using a hydraulic press ( Figure 12). To avoid eccentricities, a distribution plate was placed in the upper area of each tube. The nomenclature used for each test was as follows: H-F (hollow-flat) position, 80-115-194 mm (three tube diameters), and 5-10-20 mm/min (three speeds) [10,38]. The most important problem was the compaction of the soil each time a penetration was carried out. To solve this problem, the sand was decompressed using a kneading machine to release the sand particles (Figure 12b) [39].

Results
The results obtained in the 18 tests are shown below. First, a comparison of the three tubes is shown for each velocity and type (H or F) (Figures 13-15).   Next, a comparison of the three speeds is shown for each tube and type (H or F) (Figures 16-18).   These tests were conducted to obtain the penetration force of the driving length. Accordingly, we selected the tube, the tube position, and the penetration speed. The results are shown above.
It was verified that, for the same tube, a higher penetration speed would lead to greater force. On the other hand, it was verified whether a greater diameter of the tube led to a greater penetration force. These trials allowed the verification of both initial hypotheses [35,40].
The maximum value of penetration force obtained was 28.16 kN for the case with 194 mm diameter in a flat position, and the minimum value obtained was 1.59 kN for the case with 80 mm diameter in a hollow position.
A large amount of data were obtained from each test gauge (108 measurements). Figure 19 shows the most representative values of all measurements corresponding to Gauges 5 and 6 in the flat test. The hollow test presented smaller values than the flat test.

Discussion
The results obtained across 325 mm of deformation are recapitulated in Tables 5 and  6. These tables show the variation for each tube and each type.  Tables 5 and 6 and Figure 20 show that there is a relationship among the penetration speed, the obtained load, and the tube diameter. Taking the smallest value obtained as 1, the relative increase could be determined for the other tubes or speeds.

Conclusions
The graph in Figure 20 shows that, for the same tube, a greater speed required a greater force to achieve the same penetration [28]. For example, for a tube in a flat position with 80 mm diameter and a speed of 5 mm/min, a force of 4.48 kN was obtained; in contrast, at a speed of 20 mm/min, a force of 6.71 kN was obtained.
For the same tube, the variation of load application defined the behavior of the penetration force for the same deformation. For example, for an 80 mm tube in a flat position, there was a 1.5 ratio between the speeds of 20 mm/min and 5 mm/min. This relationship held for all tubes.
In addition, there were differences in the results obtained for the hollow and flat situations. For flat penetration, a greater force was required to achieve the same drive length, which is a vital factor to take into account when designing offshore structures.
In terms of diameter, the force/diameter ratio for tubes 1 and 3 was 4/2.4 = 1.67. This ratio is important for the estimation of penetration force values for larger-diameter tubes such as those used in offshore foundations.

Conflicts of Interest:
The authors declare no conflict of interest.