A new polyurethane-steel honeycomb composite pier anti-collision device: Concept and compressive behavior

Abstract

A new type of pier anti-collision composite structure composed of honeycomb steel and polyurethane (PU) elastomer was proposed in this study. The impacts of the shape and filling materials of inner core cells on the failure mode, load–displacement cure, bearing capacity, structural stability, and energy absorption were studied by conducting uniaxial compression tests on device segments. Test results showed that the bearing capacity, structural stability, and energy absorption of honeycomb steel structure were significantly improved by PU elastomer filling. Besides, when compared with the square honeycomb structure and the regular hexagon honeycomb structure, the maximum values of average load, total energy absorption (TEA), and specific energy absorption (SEA), which were 69.6 kN, 1986.1 J, and 1300 J/kg, respectively, for the regular triangle honeycomb structure without PU filling, increased to 459.3%, 376.38%, and 212.5%, respectively, for the regular hexagonal core cell structure with PU filling, which was proved to be the most suitable core structure for pier anti-collision device.

Keywords

Polyurethane honeycomb structure pier anti-collision device failure mode compression capacity energy absorption

Introduction

Highway transportation system has developed rapidly in China. As of the end of 2019, there were as many as 360 million vehicles, 878,300 highway bridges, and an average daily traffic with 53,429 driving vehicles in China (Department, 2019). It is worth noting that such a huge and heavy traffic usually brings about serious traffic hazards or problems, especially in some key cities such as Guangzhou, Shanghai, and Beijing, where the complicated urban transportation with enormous traffic roads and highway interchanges. As a result, collisions between vehicles and bridge piers occurred frequently. Thus, engineers and researchers have conducted a lot of research and developed various protection devices to effectively protect the bridge piers. For example, Zhou et al. (2020) proposed a new type of assembled ultra-high performance concrete (UHPC) collision avoidance, which could reduce the collision force between shipping vessels and bridges, prolong the collision time, and protect the vessels as well. Wang and Morgenthal (2019) studied an independent protective structure supported by concrete-filled steel pipes, which could protect piers from barge collision by absorbing the energy through plastic deformation. Another new vessel collision protection device made of drilled shafts, which was designed by Pinto et al. (2018), could absorb vessel collision energy through the plastic deformation of shafts. Wang et al. (2018) explored a glass fiber–reinforced plastic floating device to absorb most of the collision energy through the plastic deformation of glass fiber composite material. Wu et al. (2009) proposed an energy-consumed collision prevention system of long-distance anchor moving, which mainly used gravity anchors to consume the kinetic energy of vessels in long-distance anchors moving and prevented the vessels from hitting the non-navigational bridge. Wang et al. (2008) studied a floating device made of hundreds of wire rope coils in parallel connection to consume most of the collision energy by steering the vessels’ direction. However, it is not difficult to find that these aforementioned studies mostly focus on vessel collision of water bridge which promises sufficient space for engineers to set up large, solid protective devices in resistance to the collision from thousands or over 10 thousands of tons of vessels. But for a collision in the highway bridge, the space around piers is limited as a result, a small-sized road bridge pier protection device which can effectively absorb collision energy is needed to protect vehicles as well as the inside drivers and passengers in the collision.

The polyurethane (PU) elastomer, a linear segmented blocked copolymer, has a macromolecular main chain consisting of alternating soft (A) and hard (B) segments (Nishiyama et al., 2020; Ziegler et al., 2018), where segment A is a polyether or polyester polyol with a molecular mass of 500–3000, and segment B is generated through the reaction of isocyanate and a small molecule chain extender (Yilgör et al., 2015; Ziegler et al., 2018). Polyurethane generally has advantages of low density, high strength, good ductility, and superior energy absorption capability. When constructing structures such as vessels and buildings, sandwich composite plates filled with PU have been often used to replace pure metal plates (Momčilović and Motok, 2009). However, traditional PU sandwich composite plate does not apply to highway bridge pier collision prevention devices because of its big thickness. Therefore, it is necessary to redesign the internal structure of the PU sandwich composite plate to enhance its energy absorption. Inspired by the honeycombs which have the characteristics of good specific stiffness, buffering capacity, structural stability, specific strength, and energy absorption, the honeycomb structure has been widely applied in construction, military and aviation, etc (Nishiyama et al., 2020; Pan et al., 2006; Wang and Yang, 2000; Wierzbicki, 1983; Wu and Jiang, 1997; Kim and Christensen, 2000 Yasui, 2000; Zhao and Gary, 1998). Based on this, by combining the PU with the honeycomb structure, a new type of PU -steel honeycomb structure for pier energy-dissipating collision prevention device, which was mainly composed of outer steel plate, inner steel plate, honeycomb inner core and the filling material of PU elastomer, as shown in Figure 1, was proposed in this paper. This device can effectively absorb collision energy through internal filling PU and honeycomb core deformation to protect drivers and passengers as well as the bridge pier.

Figure 1.

Concept of pier anti-collision device.

Review studies on PU sandwich plates mainly focused on traditional sandwich structures rather than honeycomb core structures (Harris, 2007; Feldmann et al., 2007; Martin, 2007); thus, it is necessary to study the mechanical properties of PU -honeycomb core composite structure. In this paper, by conducting two sets of 18-segment scale model static compression tests, the effect of PU filling on the compression performance and energy absorption of the honeycomb core structure were studied, and the effect of core shapes on the structure was compared, so as to provide theoretical and technical supports for further research on the mechanical performance under real vehicle collision and further design optimization for this device.

Test program

Specimens and test setup

A total of 18 specimens with the size of 150 mm × 150 mm × 50 mm were prepared in this test, and the thickness of the outer steel plate was 1.5 mm. Specimens were divided into three types according to their inner core cell shapes: regular hexagon (10 mm in side length, 1 mm in cell wall thickness, and 1.568 kg of total mass), regular triangle (31 mm in side length, 1 mm in cell wall thickness, and 1.572 kg of total mass), and square(19 mm in side length, 1 mm in cell wall thickness, and 1.578 kg of total mass), as illustrated in the Figure 2(a) and Table 1. To ensure the comparability of test results, inner core size was determined in terms of the same quantity of steel used in three shapes. All specimens were divided into two groups: PU -steel honeycomb composite structure (PSHCS) and steel honeycomb composite structure (SHCS). According to their inner core shape, each group was further divided into three subgroups consisting of three identical specimens. Each Specimen was then numbered in a sequence of inner core shape—filling material—serial number in each set, as tabulated in Table 1, where H denotes the regular hexagon inner core cell, T refers to the regular triangle inner core cell, S means the square inner core cell, and P represents with PU filling. For example, H-1 represents the first regular hexagonal core specimen filled with no PU; H-P-1 refers to the first regular hexagonal core specimen with PU filled in.

Figure 2.

Specimens (unit: mm). (a) Configuration; (b) working process.

Table 1.

Specimens details.

Group	Set	Number	Cell shape	Outside size (mm)			t_s (mm)	l_c (mm)	t_c (mm)	Mass (kg)
Group	Set	Number	Cell shape	l	w	h	t_s (mm)	l_c (mm)	t_c (mm)	Mass (kg)
SHCS	H	H-1	H	150	150	50	1.5	10	1.0	1.568
		H-2
		H-3
	T	T-1	T					31		1.572
		T-2
		T-3
	S	S-1	S					19		1.578
		S-2
		S-3
PSHCS	H-P	H-P-1	H					10		2.613
		H-P-2
		H-P-3
	T-P	T-P-1	T					31		2.613
		T-P-2
		T-P-3
	S-P	S-P-1	S					19		2.614
		S-P-2
		S-P-3

Note. H is the regular hexagon; T is the regular triangle; S is the square; t_s is the thickness of outside steel plate; t_c is the thickness of cell wall; l_c is the side length of the cell section.

The steel members of each specimen were cast by stamping. Take the regular hexagonal core cell specimen shown in Figure 2(b) as an example, first, steel plate was cut into four sizes of 150 mm × 150 mm (A), 150 mm × 50 mm (B), 60 mm × 150 mm (C), and 150 mm × 10 mm (D), respectively, following the components requirement of outer top plate, outer bottom plate, outer edge plate, inner core basic layer, and inner core cell edge. Second, a bending machine was used to punch steel plate C into a corrugated plate (E) with a 120-degree bending angle and a 10 mm edge distance. Plates D and E were welded by an electric welding machine to form inner cells (F). After several Fs were assembled into a complete internal structure, plates A and B were eventually welded on their peripheral edges to produce steel members.

Polyurethane elastomer was prepared by an addition reaction of polyol containing multiple hydroxyland polyisocyanate, which has a strong positive charge distribution due to the carbon atoms in isocyanate group forming a double bond with two highly electronegative atoms, nitrogen and oxygen (Amjed et al., 2020; Christenson et al., 2007; Akindoyo et al., 2016; Xie et al., 2019). The mixing of two provided opportunity for the oxygen in polyol hydroxyl group to attack the carbon in isocyanate group, resulting in the hydrogen left for connecting with the nitrogen atom to complete the nucleophilic addition; If more than two hydroxyl or isocyanate groups were involved, addition reaction would occur continuously to finish the addition polymerization until PU was finally obtained, as shown in Figure 3(a). The entire reaction process was generally divided into two steps (Petrović and Ferguson, 1991), as shown in Figure 3(b): first, isocyanate reacted with a part of polyol to form prepolymer with a higher content of -NCO groups, free isocyanate presenting with a lower viscosity; then, it was mixed with the remaining polyol and chain extender for chain extension to produce PU products.

Figure 3.

PU elastomers (a) combination of PU; (b) chemical reaction process.

First, the plastic wrapper was employed to seal the bottom of the steel structure to prevent PU liquid from infiltrating along the bottom slits before the prepared PU elastomer was poured into the steel cells. Second, the mixed PU elastomer liquid was then injected into steel member, and the overflowing PU on the top was smoothed using a spatula. Finally, the steel structure with PU was cured at room temperature for 24 h, and the bottom plastic wrapper was removed. The whole production process is shown in Figure 2(b).

The compression test was carried out by MTS E64.106 universal testing machine using displacement loading method. The specimen was placed flat in the middle of the support, with specimen’s outer side 75 mm from the support edge, as shown in Figure 4. Then, load cell was lowered to a place 1–2 cm distance from specimen’s top surface, and pre-loading was then started at a loading rate of 5 mm/min and stopped at loading reaction reaching at 0.01 kN. After that, the operating system of the testing machine was reset to zero, and test loading started at a loading rate of 5 mm/min, which stopped when specimen was compressed to a compact state or in a pie shape. The load–displacement curve of specimen during loading process was measured by data acquisition system. Additionally, a high-definition camera was set in front of the testing machine to record simultaneously the failure process of specimen.

Figure 4.

Test setup.

Materials and properties

The steel member of the specimen was made of Q235 ordinary carbon steel plate provided by Huayuan Precision Machinery Factory. The tensile property of material samples from three groups of steel plates were tested according to the specification ISO 6892-1:2016 (ISO, 2016) using an MTSE45.305 microcomputer-controlled electronic universal testing machine. Test samples were dumbbell-shaped, the rectangular section was the test section, and the dumbbell section was the anchor end; the detailed dimensions are shown in Figure 5. Before the test, the sizes of the middle section (AA) and the two ends (BB and CC) in each sample’s rectangular section were measured with a vernier caliper, the average values of which were taken as the size of sample’s test section. The average values of the test results were taken to describe the tensile property of the steel plate. The yield strength, tensile strength, and elongation are listed in Table 2.

Figure 5.

Steel plate samples.

Table 2.

Material properties.

Data sources	Steel plate			Polyurethane
Data sources	Yield strength (MPa)	Tension strength (MPa)	Elongation (%)	Compression strength (MPa)	Tension strength (MPa)	Elongation (%)
Test results	270.7	391.3	30.7	6.2	5.5	433.5
Criterion/manufacturers	235	375–500	25	6	5.2	430

Note. The compression strength of polyurethane is the stress at the end point of the linear elastic stage in the stress–strain curve of polyurethane.

The PU elastomer of HC-8852, produced by Hecheng City Polymer Technology Co., Ltd, Shanghai, was divided into two subtypes: material A and material B. The main component of material A was PU polyol, a yellow transparent liquid with a camphor smell. The main component of material B was PU isocyanate, a colorless transparent liquid with irritating odor. Materials A and B were first heated to 60^oC and then mixed and stirred evenly to form PU elastomer. Tensile and compressive properties were tested following specifications ISO 37:2017 and ISO 7743:2017 (ISO, 2017; ISO, 2017). As shown in Figure 6, five groups of test samples were integrally cast in a polytetrafluoroethylene mold and cured at room temperature. Each test sample was polished to ensure a smooth surface before test. The specimens included dog-bone tensile specimens and cube compressive specimens. The test was carried out by using an E45.305 computer-controlled electronic universal testing machine. The average values for compressive strength, tensile strength, and elongation are shown in Table 2.

Figure 6.

PU sample.

SHCS performances

Test process and failure modes

The loading process and the final failure modes of set H, set T, and set S are described in Figure 7(a)–(c), respectively. For the test specimens with a regular hexagonal core cell in set H, as shown in Figure 5(a), no obvious change was observed in the initial stage of loading. When the loading displacement rose to 1.88 mm, buckling appeared at both ends of the specimen and the cell walls on two sides. Meanwhile, the left and right outer steel plates were partially buckled near the lower end in an outward bulging way, indicating the specimen entered into an instability stage (as shown in the load–displacement curve, section A–B in Figure 8(b)), with an initial decrease in its bearing capacity. As loading continued, the inner core cell of the specimen gradually lost their structural stability, while the outer ends kept expanding. When loading displacement reached 9.33 mm, obvious buckling deformation emerged in all the core cells inside the specimen, which meant that the entire specimen entered into a yielding state (as shown in the load displacement curve, section B-C, in Figure 8(b)). Core cells were gradually compressed with the proceeding of loading displacement until got into a closed state, resulting in the two outer steel plates quickly buckling inward. When loading displacement climbed up to 28.38 mm, all the core cells were compressed to the closed state, and specimen was transformed into a pie and stayed in its compaction stage (as in the load–displacement curve, section C–D, in Figure 8(b)). After that, the specimen showed no obvious change with the increase of loading, and the test was terminated. In conclusion, the failure mode of the specimen was buckling failure on both sides of the core cells.

Figure 7.

Failure process of SHCS. (a) Set H; (b) set T; (c) set S.

Figure 8.

L-D behaviors of SHCS specimens. (a) L–D curves; (b) Schematic of L–D curves; (c) secant stiffness; (d) peak load and average load in yield stage.

Regular triangle honeycomb specimen in set T exhibited a similar performance with hexagon specimen at the initial loading stage, but with a better elasticity than that in set H. When loading displacement approached to 3.03 mm, core cells at the specimen bottom tended to buckle and deform at a slower rate, which indicates a higher buckling resistance of the specimen when compared with set H. It can be seen from the load–displacement curve (Figure 8(a)) that the section A–B of set T is significantly larger than that of set H, accompanying with the appearance of the second peak load reaction. Loading displacement’s ascending to 6.64 mm suggests that the specimen starts to enter the yielding stage because the specimen had an apparent outward buckling in its left and right steel plates, and most of the internal core cells deformed from their original shape, especially for the bottom ones which was compressed to a closed state, while the middle core cells remained intact. When loading displacement continued its rising to 28.52 mm, specimen’s all inner core cells were compressed to a completely compact state, resulting in the stop of loading. Specimen then showed a final failure mode of being compressed to a closed state with outward buckling on both sides of inner core cells.

In comparison with the first two sets of specimens, square honeycomb specimen (set S) presented the shortest stable stage. At the loading displacement of 1.42 mm, both side plates started a rightward buckling along center line, and inner core cell wall tilted to one side, as shown in Figure 5(c). When loading displacement reached to 8.05 mm, distinct tilting deformation was generated in all core cells, accompanying with serious rightward deflection in two outer steel plates, which denotes the total yielding of specimen. Inner core cells were speedily compressed with loading to a compact state until loading displacement climbed up to 28.17 mm, when all inner core cells were compressed and compacted and loading was stopped. In general, specimen in this set exhibited a final failure mode that core cells tilted toward right along the center line until they were eventually compressed into a closed state.

Load–displacement behaviors

The compressive load–displacement curves of all three sets of specimens are displayed in Figure 8(a). Combined with the test process in Section 3.1, the load–displacement curves as shown in Figure 8(b) are divided into four sections: section O-A, section A-B, section B-C and section C-D

Section O-A shows a linear elastic load–displacement relationship, indicating that the specimens remained in elastic stage. In this stage, the inner core cell had small deformation and recovered its initial shape when unloaded. It can be seen from Figure 8(a) that area of section O-A is smaller than other three sections, and triangular core specimen had the lowest load–displacement curve slope when compared with the square core and the hexagonal core specimens, indicating that set T has the lowest stiffness in elastic stage. Figure 8(c) shows the secant stiffness of section O-A; it can be seen that set T has a secant stiffness of 40.34 × 103 kN/m, which was smaller than the 53.30 × 103 kN/m of Set H and the 105.18 × 103 kN/m of Set S. In terms of peak load, it can be seen from Figure 8(d) that set S has the maximum value of 148.3 kN, which was higher than the 100.2 kN of set H and 122.5 kN of set T. Moreover, the hexagonal specimen, which was similar to the triangular specimen, shows a similar load–displacement curve changing from a straight line to a smooth curve near the peak load.

After the peak load point A, the load reaction force of specimen decreased as deformation increased; meanwhile, significant buckling deformation occurred in inner core cells, resulting in a partial instability of the specimen until the loading arrived at point B. In this stage, set S exhibited the fastest bearing capacity decline. As indicated in Figure 8(a), when the loading displacement increases from 1.42 mm to 8.05 mm, the load reaction force dropped from the peak value of 148.3 kN–27.6 kN, which falls by 81.4%, showing that set S has the highest bearing capacity but the worst structural stability in elastic stage. In contrast, triangular core specimen (set T) and hexagonal core specimen (set H) changed much slowly, especially the set T, a secondary peak appeared in its load–displacement curve, which might be related to the partial strengthening of inner core cell on the specimen’s left top.

In section B-C, load reaction force almost has no change as displacement increased, since all the inner core cells have buckled and deformed to the closed state under continuous compression, leading to the total yielding of specimen. For honeycomb core composite structure, the overall bearing capacity was reflected by the average load reaction force F_mean at this stage (Lu et al., 2018; Ma et al., 2020; Xu et al., 2018). As shown in Figure 8(a), set T had the highest yield platform among the three sets, with the average load reaction force of 69.6 kN (Figure 8(d)), 52.9% and 102.9% higher than the other two sets (set H and set S, respectively, but procured the lowest unstablility platform, with an average load reaction force of only 34.3 kN (Figure 8(b)). Therefore, it can be concluded that triangular core specimen has a higher bearing capacity and overall stability when compared with the other two types of specimens.

Under the circumstance of loading reaching at point C, all inner core cells were compressed to the closed state, resulting in a pie-shaped and nearly solid material-structured specimen. At this stage, all specimens presented similar load–displacement relationship, that is, load reaction force strengthened rapidly with the increase of loading displacement. However, since honeycomb sandwich composite structure at this stage has completely transformed from its original form into a steel-plate–like solid material, the load–displacement in section C-D has little significance for the purpose of preventing automobile collisions (Thomas and Tiwari, 2019; Wang et al., 2020).

Energy absorption performance

The energy absorption capacity of honeycomb steel structure in this study was measured by four indexes (Baroutaji et al., 2014; Goyal et al., 2019; Guler et al., 2010; Hála et al., 2020; Hu et al., 2019; Li et al., 2018, 2020; Supian et al., 2020; Zang et al., 2020): crushing force efficiency (CFE), stroke efficiency (SE), total energy absorption (TEA), and specific energy absorption (SEA), where CFE is the ratio of average load F_mean to peak load F_max, as shown in equation (1), where the average load refers to the TEA divided by effective compression deformation (not considering the compact section of C-D). A larger CFE value corresponds to a higher structural stability during compression

CFE = F_{mean} / F_{\max} = \frac{T E A}{D_{e}} / F_{\max}

(1)

SE is the ratio of effective compression deformation to specimen height, representing the structural deformability during energy absorption, as shown in equation (2)

SE = D_{e} / h

(2)

TEA is the total energy absorption of the structure during loading process, equaling to the area enclosed by the effective deformation internal load–displacement curve and the coordinate axis, as calculated by equation (3)

TEA = \int_{0}^{D_{e}} F (x) d x

(3)

where x is the loading displacement and F(x) is the load reaction force when loading displacement equals to x. SEA is the ratio of TEA to the structure’s mass m, reflecting the energy absorption capacity per unit mass. It is generally used as an important index to evaluate structure’s energy absorption performance and calculated by equation (4)

SEA = TEA / m

(4)

The comparison diagrams of CFE, SE, TEA, and SEA for the three sets of specimens are displayed in Figure 9(a)–(d) unfold, respectively. As shown in Figure 9(a), set T had CFE values of about 0.568, 25.1%, and 145.9%, which were higher than that of set H and S, indicating that square core specimen had a high peak load, but a quickly decayed bearing capacity and a poor structural stability, while regular triangle core specimen had good structural stability during energy absorption. It can be seen from Figure 9(b) that three honeycomb steel structures have similar SE values of 0.57, 0.568, and 0.563 for set H, set T, and set S, respectively, indicating a similar deformability of three honeycomb steel structures in energy absorption under compression.

Figure 9.

Energy absorption performance of SHCS specimens. (a) CFE; (b) SE; (c) TEA; (d) SEA.

Figure 9(c) shows the TEA for specimens of the same height; it can be seen that set T has the highest TEA value of 1986.1 J, set H has a TEA value of 1292.3 J, and set S has the lowest TEA value of 967.4 J. Figure 9(d) shows the SEA for specimens of same steel quality; it can be seen that set T has the highest SEA value of 1300 J/kg, set H has the SEA value of 800 J/kg, and set S has the lowest SEA value of only 600 J/kg. It is obvious that regular triangle core specimen has a better performance in both TEA capacity and energy absorption efficiency than the other two.

PSHCS performances

Test process and failure modes

Figures 10(a)–(c) provide the failure processes and final failure modes of PSHCS specimens with the three cores of regular hexagon, regular triangle and square, respectively. For specimen with a regular hexagonal core, set H-P behaved similarly in the initial loading stage with set H (without filler). However, when loading displacement climbed about 3.29 mm, buckling deformation occurred at the inner core cells on two sides of the specimen of set H-P and the outer side plates. Meanwhile, the bottom honeycomb core cells still maintained their original shape, and the filled PU, which sustained the specimen with its bearing capacity and the load reaction force increased with the increase of loading displacement, indicating the specimen entering into the strengthening stage, as presented in Figure 11(a). As loading increased, specimen’s steel cell walls gradually buckled, as a result, the pressure load transmitted from unstable steel member to filled PU. When loading displacement reached to 9.06 mm, all the steel inner core cell walls virtually lost their stability. Some core cells were degummed with the inside PU due to large deformation, resulting in the crush of partial cell walls and the extrusion of some PU out of the cell. As loading displacement ascended to 26.06 mm, nearly all the cell walls were broken, and most of the PUs close to the specimen’s two sides were squeezed out, leaving only a small amount of PU staying in the middle cells. Specimen was thus damaged, and test was terminated after a little more displacement loaded, exhibiting a final failure mode that all the core cells and PU were degummed, in addition to the crush of steel plates on both sides.

Figure 10.

Failure process of SHCS. (a) Set H-P; (b) set T-P; (c) set S-P.

Figure 11.

L–D curves and configuration. (a) L–D curves; (b) Type I; (c) Type II.

The effect of PU filling on triangular core specimen is similar to that on the regular hexagon specimen. At the initial stage of loading, specimen showed no significant change. When loading displacement reached to 3.85 mm, buckling occurred in part of the steel cell walls while specimen bearing capacity was still strengthened due to the filling of PU, indicating that the specimen started to enter the strengthening stage. When loading displacement increased to 9.67 mm, degumming appeared, with all steel cell walls obviously buckling and fracture failure occurring at the bottom of the steel plate on specimen’s right side. When loading displacement ascended to 21.01 mm, all the core cells were squashed, provoking most PU overflowing out of the cell. Meanwhile, compared with regular hexagon and regular triangle specimens, square specimen was not significantly reinforced with PU filling. Similarly to the unfilled specimen in set S, when the loading displacement of set S-P reached to 1.62 mm, loading force dropped significantly, as shown in Figure 10(a), because less cells maintained their original shape and PU in cells remained intact but weakly restrained. When loading displacement reach to 2.91 mm, the whole specimen yielded, accompanied by the occurrence of obvious buckling deformation in all the steel cell walls and the degumming of some PU. In the later stage of the test, as loading displacement increased to 23.38 mm, all the cell walls significantly crooked. When specimen finally broken, it was folded to the right, and its upper and lower steel plates in addition to its left and right steel plates were all broken, but no large amount of PU was extruded.

Load–deformation behaviors

The load–deformation curve of PSHCS specimen is shown in Figure 11(a). Specimens were divided into two types of Type I and Type II (Figure 11(b)) by the shape of their inner core cells. The load–displacement curves of set H-P and set T-P are Type I, the rising section, which is different from unfilled specimens, consists of linear elastic section O-A and strengthening section A-B. Section O-A refers to the initial stage of loading, which is related with displacement linearly. In this stage, specimen’s steel cell walls and its internal PU, maintained their deformation within the elastic deformation range, assuring the capability of restoring back to their original state after unloading. Moreover, thanks to the less deformation of specimen, steel cell walls undertook most of the load. As loading rose to point A, part of the steel cell walls started to buckle, reducing specimen’s stiffness. Whereas, unlike that in unfilled specimen, PU prevented the specimen’s bearing capacity from descending and partook the excessive pressure load transmitted by yielding steel member. In addition, for specimen filled with PU, due to the supporting effect of PU in steel members, the load reaction force at point A was greater than the peak load of unfilled specimen. For example, set T-P increased by approximately 71.4 kN, and set H-P increased by approximately 31.8 kN, as shown in Figure 12. In addition, Figure 12(a) proves that triangular core filled specimen exhibits a higher value of point A (219.7 kN) than that of the hexagonal specimen (132 kN).

Figure 12.

Load force and improvement. (a) Peak and average value; (b) difference between SHCS and PSHCS; (c) relative improvement to SHCS.

The buckling of the sample cell wall increases with the increase of load, which makes the PU bear more and more pressure loads. When loading reached at point B, all steel members buckled, and PU is filled to bear the additional pressure load. However, meanwhile, the steel cell wall began to degum from the PU, which weakened the confinement of PU, thus slightly reducing the bearing capacity of the specimen. Then, the cells were continuously compacted, and the PU bears all initial pressure loads. The specimen thus had a slight increase in load-bearing capacity due to the supporting effect of PU, and an approximately straight-line load–displacement curve at this stage due to the elastic behavior of PU. For set HP and set TP, their load reaction force at the overall yield stage (B--C) was also greater than that of the corresponding unfilled specimens, and there was only a slight difference between the peak load and the average load reaction force, as shown in Figure 12(a), indicating that PU filling improves the structural stability. When load–displacement curve moved on to point C, all the core cells were compacted and the inside PU was squeezed out, suggesting the specimen entering into the compacting stage, where load reaction force increased rapidly with deformation.

The load–displacement curve of square core PSHCS specimen in Figure 11(c) is a Type II curve, which displays a less obvious effect of PU filling on the rising sections of the load–displacement curve, compared with that of the triangular and hexagonal core specimens. Similar to the unfilled specimen in set S, there was no strengthening section in the section O-A of the load–displacement curve for set S-P, and the load reaction force basically maintained a linear relationship with deformation before the peak value (point A). However, thanks to the supporting effect of inside PU, the peak load of set S-P was higher than that of set S. In addition, it can be clearly seen from Figure 11(a) that, same as SHCS specimen, specimen in set SP presented a shifting mode of load–displacement of first dropping suddenly to point B and then fluctuating slightly around a constant value until loading reaching to point C, different from that in sets HP and TP which increased slowly after a little decline, although the secant stiffness of set SP at this stage was greater than that of the other two sets of specimens (set HP and Set TP). The specimen in set S-P yielded overall at this stage (B--C), with its average load reaction force much smaller than that in set T-P and set H-P, but greater than SHCS specimen in set S. Similarly, after point C, all the core cells were compacted and most of the PU was extruded from the cell, suggesting the specimen entering into the compaction stage, where the specimen presented the characteristics of solid materials, and the load reaction force began to increase rapidly with deformation.

Figure 12(a) shows the peak load and the average load of PSHCS specimens with three core shapes. It can be seen that, different from SHCS specimen, triangular core specimen (set TP) had the largest peak load, 33.7 kN and 151.2 kN larger than set HP and set SP, respectively, but held an average load reaction force at the instability stage close to set HP, showing that after PU filling, specimens of regular triangle inner core and regular hexagon inner core are of little difference in their stability and that PU has a better working effect in hexagonal inner core specimen. In addition, compared with SHCS specimen, hexagonal specimen with PU filling, increased its peak load and average load by 191.4 kN (191%) and 209 kN (459.3%), respectively, as shown in Figure 12(b) and (c); triangle specimen mounted its peak load and average load by 207.3 kN (169.2%) and 189.3 kN (271.9%), respectively; square specimen leveled up its peak load and average load by 30.3 kN (20.4%) and 126.4 kN (368.5%), respectively. Therefore, it should be concluded that PU can play a better role in hexagonal core specimen.

Energy absorption performance

The values of CFE, SE, TEA, and SEA calculated by equations (1)–(4) of the three PSHCS specimens are specified in Figure 13(a)–(d), respectively. It can be seen from Figure 13(a) that, unlike SHCS specimen, hexagonal core specimen in set H-P arrayed a higher CEF value (0.859) than triangular core specimen in set T-P (0.785), indicating that hexagonal core specimen during energy absorption after being filled with PU has a better stability than triangular specimen. In terms of SE value, as Figure 13(b) shows, set H-P had the maximum value of 0.521, followed by set S-P of 0.468 and set T-P of the least 0.42, verifying that set H-P has a broader variability and a higher SE than the other two specimens.

Figure 13.

Energy absorption performance of PSHCS specimens. (a) CFE; (b) SE; (c) TEA; (d) SEA.

As for the TEA capacity, Figure 13(c) suggests that during the entire loading process, specimen in set HP exhibited the largest TEA capacity of 6632.6 J, followed by set TP of 5438.7 J and set SP of the least. Similarly, the specimen in set HP displayed the maximum TEA capacity per unit mass. As shown in Figure 13(d), set HP manifested the largest SEA capacity of 2500 J/kg, followed by set TP to 2100 J/kg and set SP minimum to 1400 J/kg. Therefore, under the same structural quality, the hexagonal core PSHCS specimen possesses the largest energy absorption capacity.

Comparison with SHCS

In order to compare the impact of PU filling on the basic properties of core structures with different shapes in detail, Figure 14(a) clarifies the increase rate of CFE, SE, TEA, and SEA values of three core shapes after filling.

Figure 14.

Relative improvement to SHCS. (a) Energy absorption performance; (b) costs.

In terms of CFE, PU filled regular hexagonal core specimen increased by about 89.21%, from 0.454 to 0.859; triangular core specimen rose from 0.568 to 0.785, with an increase rate of about 38.2%; square core specimen soared from 0.231 to 0.899, with the highest increase rate of about 289.18%. It can be seen thatcompared with specimen in SHCS, the CFE value of PSHCS specimen is greatly increased after PU filling, indicating that the structural stability of honeycomb sandwich structure during energy absorption process has been significantly enhanced after PU filling. In addition, it is clear that before the filling of PU, the triangular core specimen procured the highest CFE value among three core specimens. But after filling, the regular hexagonal core specimen assumed the largest one. This shows that PU filling can effectively improve the structural stability of the regular hexagonal cell structure to the highest. In contrast to CFE enhancement, the influence of PU filling on the SE of the specimen is weakened. From Figure 14(a), it can be seen that SE value of the hexagon core specimen after filling was reduced by 8.6%, the triangular core specimen was reduced by 26.06% and the square core sample increases by 16.87%. Therefore, since regular hexagonal core specimen obtains the least SE reduction, it has the highest SE value among the three core specimens.

Although the deformation capacity of specimens were weakened after PU filling, their energy absorption capacity was significantly improved. As Figure 14(a) shows, after filling with PU, the TEA of the hexagon core specimen increased from 1392.3 J to 6632.6 J (increased by 376.38%); The TEA of regular triangular core specimen increased from 1986.1 J to 5438.7 J (increased by 173.84%); and square core specimen alleviated it from 967.4 J to 3756.5 J (increased by 288.31%). Meanwhile, different from the triangle core specimen which possessed the largest TEA among the unfilled specimens, the regular hexagoncore specimen after PU filling assumed the largest TEA; therefore, it had the largest energy absorption capacity. As for SEA, PU has the same enhancement effect on specimens. For example, the SEA value was increased from 800 J/kg to 2500 J/kg (an increase of about 212.5%) in regular hexagonal core specimen, from 1300 J/kg to 2100 J/kg (an increase of about 61.54%) in regular triangular inner core specimen, and from 600 J/kg to 1400 J/kg (an increase of about 133.33%) in square core specimen. It thus can be seen that PSHCS specimen with a regular hexagonal core has the largest energy absorption capacity per unit mass.

In addition, the total cost and unit cost of each specimen before and after filling are shown in Figure 15(a) and (b), respectively. It is apparent that although PU filling augmented the total cost of the multi-shaped inner core specimens, when compared to that of the steel parts (about 255–340 RMB/kg), the lower unit price of PU (about 30 RMB/kg) did not lead to significant increment to the total cost of specimens, while contrarily greatly abated the unit cost of specimens. Among them, total cost increment of regular hexagonal specimens from 536 RMB to 586 RMB was the smallest, only with an increase of 9.3% (Figure 14(b)), with its unit price decrement of 34.39%, decreasing from 340.84 RMB/kg to 224.26 RMB/kg (Figure 14(b)). Total cost of regular triangle specimens increased the most from 396 ¥ to 446 ¥, with an increase of approximately 12.6%, whereas its unit price decreased from 251.91 ¥/kg to 170.69 ¥/kg (a decrease of approximately 32.34%). In addition, PU filling significantly improved the energy absorption capacity. Taking TEA as an example, specimen with a regular hexagonal core increased its TEA by about 376.38% after PU filling. In the comparison of all three inner core structures, although regular hexagonal inner core specimen cost more than triangular and square specimens, its excellent energy absorption capacity after PU filling would still make it the best option for pier anti-collision device.

Figure 15.

Costs of specimens. (a) Total costs; (b) unit costs.

Conclusion

To study the compression performance and energy absorption capacity of three shapes of inner core SHCS and PSHCS specimens, uniaxial compression tests were conducted in this paper, and the impacts of inner core shape and filler on the failure mode, load–displacement curve and the energy absorption of honeycomb structure were analyzed. The research results are as follows:

(1) For unfilled (SHCS) specimens, the regular triangle inner core structure has better structural stability, load efficiency, and energy absorption than the regular hexagonal inner core structure, but it is not significantly better. The square inner core structure has the highest peak load, but the structural stability, load efficiency, and energy absorption capacity are the worst, thereby unsuitable for being a core structure in bridge pier anti-collision device.

(2) After filling with polyurethane (PSHCS), specimens with regular hexagon and triangle cores were significantly enhanced in both structural stability and load-carrying capacity. The structural stability, load efficiency, and energy absorption capacity of three core specimens were significantly improved. In addition, the PU hexagonal inner core specimens show excellent structural stability, load efficiency, and energy absorption capability, indicating that it could be used as a good choice for bridge pier anti-collision device with the best buffing and energy absorption effect.

(3) Although the total cost of the PSHCS specimen is greater than that of the SHCS specimen, the cost increment is relatively small and the unit cost is greatly reduced. The total cost of these three core structures is increased by about 9.3%–12.6%, while the unit cost is reduced by about 32.2%–34.4% and the energy consumption performance is increased by 173.84%–376.38%. This shows that PU filling in this protective structure will not only significantly improve its energy absorption capacity and enhance protection, but also result in better economic benefits. In addition, although the cost of hexagonal inner core specimen is slightly higher than the other two, considering its excellent energy absorption and compression performance after PU filling, hexagonal inner core structure will still be the best selection for pier anti-collision device.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was financially supported by the National Natural Science Foundation of China [grant numbers 51608137] and Innovation Training Project of Guangzhou University [grant number 201911078036].

ORCID iD

Liwen Zhang

Junping Zhang

References

Akindoyo

Beg

MDH

Ghazali

, et al. (2016) Polyurethane types, synthesis and applications-a review. Rsc Advances 6(115): 114453–114482.

Amjed

Bhatti

Zia

, et al. (2020) Synthesis and characterization of stable and biological active chitin-based polyurethane elastomers. International Journal Of Biological Macromolecules 154: 1149–1157.

Baroutaji

Morris

Olabi

(2014) Quasi-static response and multi-objective crashworthiness optimization of oblong tube under lateral loading. Thin-Walled Structures 82: 262–277.

Christenson

Anderson

Hiltner

(2007) Biodegradation mechanisms of polyurethane elastomers. Corrosion Engineering, Science and Technology 42: 312–323.

Department

(2019) Statistical Bulletin on the Development of the Transportation Industry in 2019 Year. Available at: http://xxgk.mot.gov.cn/2020/jigou/zhghs/202006/t20200630_3321335.html.

Feldmann

Sedlacek

Geßler

(2007) A system of steel-elastomer sandwich plates for strengthening orthotropic bridge decks. Mechanics Of Composite Materials 43: 183–190.

Goyal

Anand

Sharma

, et al. (2019) Crashworthiness analysis of foam filled star shape polygon of thin-walled structure. Thin-Walled Structures 144: 106312.

Guler

Cerit

Bayram

, et al. (2010) The effect of geometrical parameters on the energy absorption characteristics of thin-walled structures under axial impact loading. International Journal of Crashworthiness 15: 377–390.

Hála

Frydrýn

Máca

, et al. (2020) Comparative study on the crashworthiness of brittle blocks with thin-walled cellular structures. Thin-Walled Structures 148: 106578.

10.

Harris

(2007) Lateral Load Distribution and Deck Design Recommendations for the Sandwich Plate System (Sps) in Bridge Applications. Virginia Polytechnic Institute and State University.

11.

Wang

Song

, et al. (2019) Energy-absorption characteristics of a bionic honeycomb tubular nested structure inspired by bamboo under axial crushing. Composites Part B: Engineering 162: 21–32.

12.

ISO 6892-1:2016 (2016) Metallic materials - Tensile testing - Part 1: Method of test at room temperature.

13.

ISO 37:2017 (2017) Rubber, Vulcanized or Thermoplastic—Determination of Tensile Stress-Strain Properties.

14.

ISO 7743:2017 (2017) Rubber, Vulcanized or Thermoplastic—Determination of Compression Stress-Strain Properties.

15.

Kim

Christensen

(2000) Basic two-dimensional core types for sandwich structures. International Journal of Mechanical Sciences 42: 657–676.

16.

Chen

(2018) Comparative analysis of crashworthiness of empty and foam-filled thin-walled tubes. Thin-Walled Structures 124: 343–349.

17.

Hou

, et al. (2020) Crashworthiness analysis of corrugations reinforced multi-cell square tubes. Thin-Walled Structures 150: 106708.

18.

Gao

, et al. (2018) Crushing analysis and crashworthiness optimization of tailor rolled tubes with variation of thickness and material properties. International Journal Of Mechanical Sciences 136: 67–84.

19.

Xie

(2020) Crashworthiness analysis of thin-walled bio-inspired multi-cell corrugated tubes under quasi-static axial loading. Engineering Structures 204: 110069.

20.

Martin

(2007). Sandwich Plate System Bridge Deck Tests. Virginia Polytechnic Institute and State University.

21.

Momčilović

Motok

(2009) Estimation of ship lightweight reduction by means of application of sandwich plate system. FME Transactions 37: 123–128.

22.

Nishiyama

Kumagai

Motokucho

, et al. (2020) Temperature-dependent pyrolysis behavior of polyurethane elastomers with different hard- and soft-segment compositions. Journal of Analytical and Applied Pyrolysis 145: 104754.

23.

Pan

Sun

, et al. (2006) Longitudinal shear strength and failure process of honeycomb cores. Composite Structures 72: 42–46.

24.

Petrović

Ferguson

(1991) Polyurethane elastomers. Progress In Polymer Science 16: 695–836.

25.

Pinto

Prato

Huerta

PJF

(2018) Vessel colission protection for chaco corrientes bridge by means of energy absorbing drilled shafts. Mecanica Comput 27: 813–832.

26.

Supian

ABM

Sapuan

Zuhri

MYM

, et al. (2020) Crashworthiness performance of hybrid kenaf/glass fiber reinforced epoxy tube on winding orientation effect under quasi-static compression load. Defence Technology 16: 1051–1061.

27.

Thomas

Tiwari

(2019) Energy absorption and in-plane crushing behavior of aluminium reinforced honeycomb. Vacuum 166: 364–369.

28.

Wang

Yang

(2000) Damping of honeycomb sandwich beams. Journal of Materials Processing Technology 105: 67–72.

29.

Wang

Hao

(2020) The quasi-static axial compressive properties and energy absorption behavior of ex-situ ordered aluminum cellular structure filled tubes. Composite Structures 239: 112039.

30.

Wang

Song

Wang

, et al. (2018) Evaluation of composite crashworthy device for pier protection against barge impact. Ocean Engineering 169: 144–158.

31.

Wang

Yang

, et al. (2008) An impact dynamics analysis on a new crashworthy device against ship–bridge collision. International Journal of Impact Engineering 35: 895–904.

32.

Wang

Morgenthal

(2019) Parametric studies of pile-supported protective structures subjected to barge impact using simplified models. Marine Structures 63: 138–152.

33.

Wierzbicki

(1983) Crushing analysis of metal honeycombs. International Journal of Impact Engineering 1: 157–174.

34.

Jiang

(1997) Axial crush of metallic honeycombs. International Journal of Impact Engineering 19: 439–456.

35.

Chen

(2009) An energy consumed collision-prevention system of long distance anchor moving for non-navigational bridge. Highway 1: 213–218 (in chinese)

36.

Xie

Zhang

Bryant

, et al. (2019) Degradation and stabilization of polyurethane elastomers. Progress In Polymer Science 90: 211–268.

37.

Zhang

Wang

, et al. (2018) Crashworthiness design of novel hierarchical hexagonal columns. Composite Structures 194: 36–48.

38.

Yasui

(2000) Dynamic axial crushing of multi-layer honeycomb panels and impact tensile behavior of the component members. International Journal of Impact Engineering 24: 659–671.

39.

Yilgör

Wilkes

(2015) Critical parameters in designing segmented polyurethanes and their effect on morphology and properties: a comprehensive review. Polymer 58: A1–A36.

40.

Zang

Zhang

, et al. (2020) Crashworthiness of CFRP/aluminum alloy hybrid tubes under quasi-static axial crushing. Journal of Materials Research and Technology 9: 7740–7753.

41.

Zhao

Gary

(1998) Crushing behaviour of aluminium honeycombs under impact loading. International Journal of Impact Engineering 21: 827–836.

42.

Zhou

Wei

, et al. (2020) Design and simulation analysis of a new type of assembled UHPC collision avoidance. Applied Sciences 10(13): 4555.

43.

Ziegler

Guttmann

Kopeinig

, et al. (2018) Influence of different polyol segments on the crystallisation behavior of polyurethane elastomers measured with DSC and DMA experiments. Polymer Testing 71: 18–26.