Effect of BoltHole Clearance on Bolted Connection Behavior for Pultruded Fiber-Reinforced Polymer Structural Plastic Members

Bolt-hole clearance affects the failure mode on the bolted connection system of pultruded fiber-reinforced polymer plastic (PFRP) members. The various geometric parameters, such as the shape and cross-sectional area of the structural members, commonly reported in many references were used to validate the bolt-hole clearance. This study investigates the effects of the bolt-hole clearance in single-bolt connections of PFRP structural members. Single-bolt connection tests were planned using different bolthole clearances (e.g., tight-fit and clearances of 0.5mm to 3.0mm with 0.5mm intervals) and uniaxial tension is applied on the test specimens. Most of the specimens failed in two sequential failure modes: bearing failure occurred and the shear-out failure followed. Test results on the bolt-hole clearances are compared with results in the previous research.


Introduction
Until the 1990s, the use of fiber-reinforced polymer plastic (PFRP) composites was limited to aerospace and military applications [1].However, the fiber-reinforced plastic (FRP) composites have many advantageous mechanical properties, such as an excellent strength-to-weight ratio and stiffnessto-weight ratio, which make them highly desirable also as a building material for civil engineering applications [2,3].Therefore, efforts to include FRP materials in civil engineering have been increased markedly over recent decades.In order to use FRP materials in the construction field, the pultruded structural member must be connected.Several types of connections are currently used for this purpose, including bolted, bonded, a combination of bolted and bonded, and interlocking connections.For civil engineering applications, bolted connections are preferred and deemed the most practical because they are easy to assemble and disassemble, easy to maintain, and are usually costeffective when compared with the other types of connections [4].
In general, there are four possible failure modes for pultruded fiber-reinforced polymer plastic (PFRP) singlebolted connections that are subjected to tensile forces [5][6][7][8][9][10][11][12], as shown in Figure 1. Figure 1(a) illustrates net-tension failure, which is attributable to the reduced cross-sectional area of a FRP member that is due to the bolt hole.Figure 1(b) shows cleavage-tension failure where the cross-sectional area of the bolt is not resistant to tensile loading and breaks away from the contact point.Figures 1(c) and 1(d) show bearing failure and shear-out failure, respectively.
Conventionally, FRP materials are failed in a brittle mode [12], whereas bearing failure allows ductile behavior unlike the other failure modes (net-tension failure, cleavage-tension failure, and shear-out failure).Bearing failure is the most likely to occur when FRP material is used in a structural member.Various parameters need to be taken into account to ensure safety in the design with regard to the bolt connection in order to induce bearing failure.Researchers and manufacturers who have studied bolt connections of composite materials have suggested a geometric recommendation index based on experiments and experience [16][17][18][19][20][21][22][23][24].Geometric coefficients for a bolt's width-to-diameter ratio of 6 (i.e., / = 6) and end-to-diameter ratio of 3 (i.e., / = 3) were proposed to induce bearing failure according to ASTM Standard D5961/D5961 M-96 [21].Rosner and Rizkalla [6] reported that bearing failure mode occurs predominantly with an increase in the use of high geometric coefficient values for both the width-to-diameter ratio and end-to-diameter ratio.They recommended that the geometric coefficients (/ and /) should be greater than 5.0.Therefore, the geometric coefficients used in this study are / = 5 and / = 5.
The clearance between the bolt diameter (  ) and the hole diameter ( ℎ ) is one of the parameters that allows constructability and ductile failure in the bolt connection.However, very few experiments have been conducted with regard to the necessary clearance, and existing codes recommend various values to determine the appropriate bolt-hole clearance.Thus, this study aims to address this informational deficit and determine the adequate bolt-hole clearance to prevent brittle failure in a single-bolted connection system for PFRP materials.A total of 98 single-bolt connection specimens with various bolt-hole clearances were tested under tensile loading.Also, the existing available code values found in the Eurocomp Design Code [13] and the Italian National Research Council (CNR) standards [14] and the bolt-hole clearance reported by Mottram [15] were compared with the experimental results obtained.

Experimental Program
2.1.Mechanical Properties of PFRP Structural Members.Two shapes, angle, and I-shape, of PFRP structural members were investigated.The PFRP structural members were fabricated in Korea using the pultrusion process and manufactured using E-glass fiber and polyester resin with the fiber volume fraction of 0.578.In the pultruded structural shapes, E-glass fiber bundles are placed along the longitudinal direction of the member.
Figure 2 presents the dimensions of two types of PFRP specimens.
The material properties of the PFRP members were determined from tensile tests, compression tests, and shear tests.The tensile test specimens were taken in the longitudinal direction (i.e., the member axis direction that coincided with the reinforcing fiber direction) and prepared, with slight modification, according to ASTM D3039/D3039 M-08 [26]. Figure 3 shows the prismatic tensile test specimens and test setup.Fifteen specimens were prepared, including five specimens from the angle (specimen dimension: 250.00 × 24.78 × 9.82 mm) and ten specimens from the I-shape (specimen dimension: 250.00 × 25.44 × 9.62 mm).Each specimen was loaded up to failure with a loading speed of 3 mm/min in accordance with the displacement control method.In these tensile tests, all the specimens are failed in a brittle manner within the gage length.
In order to determine the modulus of elasticity along the transverse direction of each specimen, tensile strength test specimens should be taken along the transverse direction of the member to comply with ASTM D3410/D3410 M-03.However, in this study, the compressive strength test, instead of tensile test, suggested by Yoon [27] was adopted because taking a tension test specimen in the transverse direction of the member [28] is not possible because the width of flange and web are too small.Figure 4 shows the compression specimens and test setup used in this study.Fifteen specimens were prepared, including five specimens from the angle of structural member (specimen dimension: 80.00 × 19.79 × 9.83 mm) and ten specimens from the I-shape (specimen dimension: 80.00 × 18.29 × 9.50 mm).Each specimen was loaded up to failure with a loading speed of 3 mm/min  according to the displacement control method.In these compressive tests, all the specimens are failed in a brittle manner within the gage length.Finally, shear tests for the PFRP specimens were also conducted according to the method found in ASTM D5379/D5379 M-12 [29].Figure 5 shows a shear test specimen and test setup.Thirty specimens were prepared, including 10 specimens from the angle type (specimen dimension: 76.27 × 12.06 × 9.22 mm) of structural member in both the longitudinal direction and transverse directions and 20 specimens from the I-shape type (specimen dimension: 76.01 × 12.04 × 12.20 mm) also in the longitudinal and transverse directions.Load was applied with a speed of 1.27 mm/min according to the displacement control method.
Table 1 presents a summary of the average test results for the three types of tests performed in terms of the specimens' mechanical properties.were cut from two types of PFRP structural members (angle and I-shape) using a table-saw with a diamond-tipped blade.At one end of each plate, holes were marked and drilled the desired end distance according to the specified pattern for the connection.Rectangular plates for I-shape-1, I-shape-2, and I-shape-3 were cut from the flange and web of I-shape PFRP member.Table 2 provides detailed dimensions of the single-bolted test specimens.A total of 98 bolted connection specimens (i.e., angle and I-shape-3 specimens: three for each test variation, I-shape-1, and I-shape-2 specimens: four for each test variation) were prepared by drilling holes with different bolt-hole clearances.The specimens identified with a member shape, bolt diameter, and bolt-hole clearance.

Single-Bolted
Stainless steel hexagon head screws (M10) [30], stainless steel hexagon nuts (M10) [31], and stainless steel plain washers [32] were used in the fabrication of the connection test specimens that were taken from the angle and I-shape-1 structural members.Steel hexagon head bolts (M10, M12) [30], steel hexagon head nuts (M10, M12) [31], and stainless steel plain washers [32] were also used in the fabrication of the connection test specimens taken from the I-shape-2 and I-shape-3 specimens.Figures 6 and 7 describe the geometric parameters for the single-bolted connections and the three types of bolts (including nuts and washers) used in the test, respectively.
Four basic geometric parameters that may affect the strength and the failure mode of single-bolt connections were investigated in the experimental program: the width of the member (), the end distance (), the hole clearance ( ℎ −   ), and the reinforcing fiber direction of the PFRP structural shapes (refer to Figure 6).Rosner and Rizkalla [6] reported that the thickness () of the specimen barely affects the experimental results in overlapped spliced joint tests; thus, the dimensions specified by the manufacturer were used without any further process.

Tension Tests for Single-Bolted Connections.
Tension tests using single-bolted connections were conducted using a 1000-kN Universal Testing Machine. Figure 8 presents the tension test for a double-lapped joint and the test setup.The grip plates were made using a stainless steel plate.Also, a stainless steel hexagon head screw (M10) and steel hexagon head bolt (M10, M12) were used to fabricate the connection test system.Tensile loading was applied with a displacement rate of 1 mm/min (0.167 mm/sec) in accordance with ASTM D953-10 [33].

Tension Test Results for Single-Bolted Connections.
As shown in Figure 9, local failure load is defined as the load   3 presents a summary of the local failure loads, structural fracture loads, and corresponding modes of failure for all the tested connection specimens for the two structural profile sections used in this investigation.In Table 3, the failure modes are described as CT (cleavagetension), B (bearing), and S (shear-out).There is no nettension failure (NT) in this study.

Effects of Bolt-Hole Clearance.
For civil engineering applications, maintaining a uniform and precise size for the bolt-hole clearance is important for constructability.Table 4 shows the bolt-hole clearance suggested in the Eurocomp Design Code [13], the Italian National Research Council (CNR) standard [14], and previous research [15].
To investigate the effects of variation in the bolt-hole clearances, the failure loads were plotted with respect to the bolt-hole clearances ( ℎ −   ), as shown in Figures 12-15.
Figures 12-14 shows no noticeable trend in the changes in the structural fracture load with respect to the bolthole clearance.In contrast, a significant decreasing trend is evident in the results for the local failure load.Therefore, the difference between the local failure load and the structural fracture load is larger if the bolt-hole clearance has increased.However, Figure 15 shows that the structural fracture load for I-shape-3 with respect to the bolt-hole clearance has a decreasing trend.
In previous research [34], coupon tests were conducted using the whole cross-section of pultruded I-shape specimens.A total of 18 coupons were cut from the cross-section: six coupons from the upper flange, six coupons from the lower flange, and six coupons from the web.The results for the material property variation tests show that the longitudinal elastic modulus is different in the range of 9 percent to 23 percent.This variation of the elastic modulus may affect the failure load of the specimen.The reinforcing fiber direction and sampling location in the structural shape produced by the pultrusion process may result in differences between the elastic moduli of the web and flange of the PFRP structural shapes.Therefore, the structural fracture load decreased slightly.Similarly, the discrepancy between the local failure load and structural fracture load increased with an increase in the bolt-hole clearance.
A relatively large bolt-hole clearance is preferred for easy of fabrication of the structure to secure constructability.The recommended value suggested in the Eurocomp Design Code [13] seems to be too small to use in practice.In addition, the application of the Eurocomp simplified method for the design of bolted joints for PFRP materials is questionable because of its reliance on single-curve normalized stress distributions.The CNR standard [14] proposes a bolt-hole clearance of 1 mm.However, because the time required for the structural fracture load to occur after the local failure load test is too short, the test specimen is not suitable for use as a design basis in consideration to safety.Therefore, the suggested bolthole clearance of 1.6 mm (1/16 in.) found in the Mottram [15] is preferable because this clearance dimension accounts for the span between the local failure point (which is the ductile failure mode) and the structural failure point (which is the fracture point), even if the FRP is a brittle material.The ASCE Design Guide for FRP Composite Connections [24] International Journal of Polymer Science        was discussed on the effect of bolt fit (hole clearance).The hole diameter, , equals the bolt diameter,   , plus 5/8 in.(15.875 mm) which may acceptable in practice.However, it is felt that 15.875 mm hole clearance is too large to be efficient to induce bearing failure mode.

Conclusions
In this paper we investigated the effects of bolt-hole clearance in single-bolted connections in the PFRP structural members.Different sizes of bolt-hole clearance from tightfit to 3.0 mm with 0.5 mm intervals were investigated.The experimental results in terms of local failure load, structural fracture load, and failure mode were analyzed with respect to the geometric parameters (i.e., the bolt-hole clearances).The following results were found.
(1) The specimens, in general, failed with two sequential failure modes.The bearing failure mode appeared first and the shear-out failure mode followed.Therefore, the geometric parameters of the specimens, that is, the / and /, were needed to maintain with sufficient values regardless of the bolt-hole clearance.
For each case, bearing failure was the predominant failure mode.
(2) When the bolt-hole clearances were in the range of 0 mm to 3 mm, no significant trend was evident with regard to the structural fracture loads.However, the local failure load decreased if the bolt-hole clearance was increased.Differences between the structural fracture loads and the local failure loads were greater when the bolt-hole clearance was increased.
(3) Constructability can be ensured by maintaining a minimum bolt-hole clearance.The Eurocomp Design

Figure 1 :
Figure 1: Typical failure modes of bolted connections.

Figure 3 :
Figure 3: Tensile test specimens and test setup.

Figure 4 :
Figure 4: Compression test specimens and test setup.

Figure 5 :
Figure 5: Shear test specimen and test setup.
bearing failure mode changes to the shearout failure mode.The load-displacement curves for the specimens indicate a local failure load after which the load increases and then decreases slightly and continuously up to the point where the structural fracture load is reached.Figures10 and 11show the failure modes at each stage of load increment.Table

Table 1 :
Mechanical properties of PFRP members.

Table 2 :
Dimensions of test specimens connected with a single-bolt./  /    /  ℎ −   Number of specimens

Table 3 :
Experimental results for PFRP bolted connections.Shape ID /  /    /  ℎ − Note.Although all of the specimens were assumed to have exact sizes for bolt-hole clearance, however, in practice some error in clearance size was unavoidable during the specimen preparation and testing, which may affect the test results.