A Comparison of Individual Bored Pile Bearing Capacity Using the Results of Standard Penetration Test (SPT) and Pile Driving Analysis (PDA) Test of the Railway Bridge Foundation

Prediction of the bearing capacity of deep foundations using analytical methods from the results of the standard Penetration Test has been carried out until now due to practical consideration. However, with the many types of empirical formulas proposed by various researchers, it is necessary to test the validity of the values predicted by the empirical formulas. During the construction stage, the bearing capacity of deep foundations can be checked using the Pile Driving Analysis (PDA) Test. This study aimed to compare the foundation bearing capacity prediction results using analytical methods of Luciano Decourt (1982), Reese and O'Neil (1999), and Building Standard Law of Japan with the results from PDA test. The comparison indicate that the analytical method of Luciano Decourt (1982) predict bearing capacity with the smallest error rate, which is 1.97% to 38.99% with safaty factor (SF) 2 to 3 on a relatively homogeneous and heterogeneous soil conditions. Prediction of bearing capacity of bored pile foundations using analytical methods O'Neil & Reese (1999) and Building Standard Law of Japan provide better performance for soil types which is dominated by cohesive soils with safety factor (SF) 2.


INTRODUCTION
Foundations are structures that are very important in construction to transfer the load of the building or the bridge to soil. In some cases, shallow foundations are inadequate to support the structural load, and deep foundations are required [1]. Types of deep foundations are classified according to geologic condition, installation method, structural material, ground effect, function, cross-section, loading, isolation, inclination, and other characteristics [2].
Prediction of the bearing capacity of deep foundations using analytical methods from the results of the standard Penetration Test has been carried out until now due to practical consideration. However, with the many types of empirical formulas proposed by various researchers, it is necessary to test the validity of the values predicted by the empirical formulas. During the construction stage, the bearing capacity of deep foundations can be checked using the Pile Driving Analysis (PDA) Test. This study aimed to compare the foundation bearing capacity prediction results using analytical methods of Luciano Decourt (1982), Reese and O'Neil (1999), and Building Standard Law of Japan with the results from PDA test. The comparison indicate that the analytical method of Luciano Decourt (1982) predict bearing capacity with the smallest error rate, which is 1.97% to 38.99% with safaty factor (SF) 2 to 3 on a relatively homogeneous and heterogeneous soil conditions. Prediction of bearing capacity of bored pile foundations using analytical methods O'Neil & Reese (1999) and Building Standard Law of Japan provide better performance for soil types which is dominated by cohesive soils with safety factor (SF) 2. [15] Various analytical methods are developed to predict the bearing capacity of deep foundations. Although analytical methods based on laboratory test data can produce good predictions, analytical methods using empirical formulas from soil investigations data, such as the Standard Penetration Test (SPT), continue to be developed. The development of this empirical analysis method aimed to make the analysis process more practical and faster. Several studies compared the bearing capacity of the analytical method and field test [3], [4], [5], [6], [7]. However, with the development of various methods of analyzing the bearing capacity of deep foundations, it is necessary to determine which method estimates the closest value of the bearing capacity prediction data compared to the bearing capacity of the test results in the construction stage.
This study aimed to continue discussing bearing capacity estimation for two deep foundations of railway bridges in Java with differences in their soil characteristics. The first one is a railway bridge identified as BH 122 in Krian, Sidoarjo, from the Surabaya-Solo railway line, where soil characteristics are dominated by cohesive soil. The second one is named BH 05 in Kulon Progo, Yogyakarta, from Kedundang Station-Yogyakarta International Airport (YIA) railway line, where the soil characteristics contain silty clay and sand. The calculation method of bearing capacity uses the method of Luciano Decourt (1982) [8], O'Neil & Reese (1999) [9], and the Building Standard Law of Japan (Japanese Method) [5]. A comparison was made with the PDA test results to validate the bearing capacity in the construction.

N-SPT Correction
The procedure of the Standard Penetration Test (SPT) can be referred to as ASTM D1586 [10] or SNI 4153:2008 [11]. SPT can provide useful and reliable data for the geologic characterization of a site, primarily due to its simplicity and relatively low cost [12]. However, the estimation of SPT can produce inaccurate results due to environmental conditions such as the presence of groundwater levels. The raw SPT data could be improved by applying certain correction factors [13]. Therefore, Terzaghi dan Peck [14] propose a correction for the N-SPT value at points below the groundwater level and the N-SPT value > 15 with the following formula: N'-SPT = 15 + 0.5 (N-15) (1) Where: N'-SPT = correction of the number of strokes N-SPT = the number of strokes in the field

N-SPT correlation with soil parameters
N-SPT is only an index of soil behavior and does not directly measure any of the conventional engineering properties of soil [13]. However, N-SPT correlation with soil parameters is carried out to provide the data needed to complete the prediction formula for the bearing capacity of the bored pile foundation. Equations (2) to (4) are empirical equations that can be used to determine the value of soil cohesion (cu) based on the SPT test data [15]. Plastic clay, cu = 12.5 N'-SPT (kPa) (2) Silty clay, cu = 10 N'-SPT (kPa) (3) Sandy clay, cu = 6.7 N'-SPT (kPa) (4)

Analytical Method for Piles Bearing Capacity
The ultimate bearing capacity of a pile is the maximum load the pile can carry without failure or excessive settlement of the ground. It depends on the soil type, cross-section, and pile length [16]. In general, the formula in Eq. (5) and (6) is used for calculating the ultimate bearing capacity of the pile for deep foundations (Qu). According to SNI 8460-2017, a deep foundation should be designed with a safety factor of 2.5 at the minimum [17].
Where: Qp = Theoretical bearing capacity for the tip of foundation, or end bearing (Ton) [16] qp = Theoretical unit tip-bearing capacity (Ton/m 2 ) Ap = Effective area of the tip of the pile (m 2 ) Qs = Theoretical bearing capacity due to shaft friction, or adhesion between foundation shaft and soil (Ton) qs = Theoretical unit friction capacity (Ton/m 2 ) As = Effective surface area of the pile shaft (m 2 ) SF = Factor of Safety/Safety Factor Pile bearing capacity could be estimate from soil laboratory test or soil investigation result, such as SPT and CPT (Cone Penetrometer Test) [18]. The method to analyze the bearing capacity of the empirical bored pile foundation using the data from the SPT test is as follows: A. Luciano Decourt Method (1982) [15] • Stress at the end of the pile, qp qp = (N ̅ p . K) (7) Where: N ̅ p = average value of SPT around 4B above and 4B below the pile foundation, where B is the diameter of the foundation K = coefficient of soil characteristic value: 12 ton/m 2 for clay 20 ton/m 2 for clay silt 25 ton/m 2 for sandy silt 40 ton/m 2 for sand • Stress due to lateral attachment, q s qs (ton/m 2 ) = ( Where: N ̅ s = average value along the length of the pile embedded within the boundary 3 ≤ N ≤ 50 B. O'Neil & Reese Method (1999) [9] The empirical method of Reese and O'Neill (1988) is considered reliable in estimating the bearing capacity of a pile foundation based on soil data [19]. The method was explained with the following formulas.
• Bearing capacity at pile tip, Qp for cohesive soil: If foundation depth < 3B, then: [1 + 1 6 ] . N * c. su (9) If su ≥ 96 kPa and the depth of the foundation base ≥ 3B, then: Qp = 9.Su (10) If su < 96 and the depth of the foundation base ≥ 3B, then: [ln (I r + 1)]. su = N * c.su (11) Where: N * c = bearing capacity factor (  (14) Where: α = dimensionless correlation coefficient, which value is determined by: α = 0 between the soil surface to1.5 m depth or to a depth with changes in water content at a deeper location. α = 0.55 if su/pa ≤ 1.5 and varies linearly between 0.55 dan 0.45 for su/pa between1.5 and 2.5. pa = atmospheric pressure (101 kPa) • Ultimate Shaft Resistance Capacity, Qs for a non-cohesive soil qs = β.σv' (15) where: β = dimensionless correlation coefficient between effective vertical stress and qs-max for layer i. The total load capacity of bored pile is calculated using Equation (5). Stress at pile tip (qp) and friction bearing capacity (Qs) calculated using Equation (18) Where: ̅ = Average value along the embedded pile with limits ≤ 30 Ls = total length of the pile in contact with sand q ̅ u = average shear strength of clay with a value ≤ 200 kPa Lc = total length of the foundation piles in contact with the clay φ = cross-section circumference of the foundation

Characteristics of the Soil
The SPT (Standard Penetration Test) test results for the BH 122 Surabaya-Solo railway bridge in Krian, Sidoarjo, and The BH 05 Bridge Crossing Kedundang Station -Yogyakarta International Airport (YIA) in Kulon Progo, Yogyakarta is compared in Figure 1. The figure also includes the N'-SPT value for bridge BH 05, which is the N-SPT value after being corrected using the Terzaghi and Peck formula.
Based on Figure 1, the type of soil in the BH 122 bridge area in Sidoarjo is dominated by cohesive soil types, namely silt and clay. While the soil composition in the BH 05 bridge area in Yogyakarta is more heterogenic, the top layer of soil to a depth of 15 m is a cohesive soil type, and at the bottom is sandy soil. The position of the groundwater level in the abutment area of the two bridges is relatively near to the topsoil, 1 m below ground level for the BH 122 bridge and 1.8 m below the ground level for the BH 05 bridge. [18]

Bored Pile Foundation
The final depth of soil bearing capacity calculation for BH 122 and BH 05 is 39 m and 24 m depth. Calculation of the bearing capacity of the foundation in the bored piles was conducted using the analytical method from Luciano Decourt (1982), O'Neil & Reese (1999), and the Japanese method of the BH 122 and BH 05 bridges. The results of the calculation with various safety factor are shown in Table 2 and Table 3. The example of the bearing capacity estimation of bored piles from each method on the BH 122 bridge at a depth of (d) 12 m is shown in the following discussion.

A. Luciano Decourt Methods (1982)
Based on Figure 1, the type of soil at a depth of (d) 12 m of the BH 122 bridge is silt and clay, so the K value in the stress analysis at the end of the pile is 20 tons/m 2 . The N ̅ p is taken from the average N-SPT at about 4 meters (4B) above and below the pile tip location, i.e., 8 to 16 m depth. Meanwhile, N ̅ s taken from the average N'-SPT from a depth of 1 to 12 meters. The bearing capacity of the bored pile foundation is calculated using Equations (7) and (

C. Japanese Methods
The bearing capacity of the bored pile foundation using the Japanese method is determined using equations (18)

Comparison of Bearing Capacity from SPT Empirical Method and PDA Test Results
The evaluation of pile's bearing capacity by analytical methods and further checking of the result during field test is an important stage of the pile foundation design [20]. Based on the analysis of the bearing capacity of the bored pile foundation from PDA test results using CAPWAP application, the bearing capacity of the bored pile foundation of the BH 122 Bridge is 341 tonf and the bearing capacity of the bore pile foundation of the BH 05 Bridge at Yogyakarta International Airport (YIA) is 689 tonf. The bearing capacity of the foundation in bored piles using analytical methods with various assumptions of safety factors on Table 2 and Table 3 is compared to the results of the CAPWAP analysis and is shown in Table 4.