Keywords

1 Introduction

The economic growth and rapid industrialization of the past decades have led to a larger number of efficient and robust steel pallet rack systems. These structures mainly use cold-formed steel (CFS) profiles due to their high strength-to-weight ratio. At the same time, these structural systems can be customized and are easy to assemble in various configurations. The slenderness and method of cold forming led to inherent vulnerabilities, especially in the uprights at the base of the structures. In addition, the almost complete lack of structural redundancy makes these elements essential for the overall stability and load-bearing capacity of the entire system. There are numerous research studies and publications, especially in recent years [1,2,3] and a considerable number of design codes [4,5,6,7].

Experimental studies on the behavior of cold-formed profiles and finite element numerical analyses have been conducted in various research centers. For example, in the papers [8, 9] the authors proposed an efficient method to increase the resistance capacity of vertical frames under compression using bolts and spacers distributed in intervals along the height. They conducted experimental tests on 81 vertical frames with different thicknesses and heights, and the effect of employing strategies to improve structural performance was examined through failure mechanisms and load-bearing capacity. The study concluded that the effect of section thickness on the performance of vertical elements is considerable and should be investigated through experimental tests. It was also observed that the structural performance of longer strengthened or non-strengthened specimens is not influenced as significantly as shorter specimens. This is mainly because the buckling failure of longer uprights is governed by the flexural and torsional flexural buckling, and strengthening solutions does not necessarily change or control the buckling mode.

In particular, the vulnerability of the base upright profiles has been the focus of recent works such as [10,11,12], where the failure modes under axial loading were discussed in detail. The paper [12] focused on the experimental investigation of interactive buckling in steel pallet rack compressed members. The study was carried out at the CEMSIG Research Centre of the Politehnica University of Timisoara, where both perforated and non-perforated upright members with two different cross-sections were tested according to the European design code EN15512 for pallet rack systems. The aim was to determine the ultimate strength of the specimens corresponding to local and distortional buckling for critical length equal to the distance between two subsequent nodes of an upright frame. Special attention was paid to the observation of distortional buckling in upright members. It was found that due to varying cross-sectional dimensions, the length between nodes often exceeded the distortional critical length, leading to results that correspond more to the interaction between distortional and global buckling rather than pure distortion. To address this, additional specimens were tested at lengths calibrated for distortional buckling and interactive distortional-overall buckling range.

In a context where there are already stores and warehouses with pallet rack storage systems (as can be seen in Fig. 1) which are nearly 30 years old, built according to the standards available at that time, coupled with the shock from the recent seismic activities in 2023 in the world, there are serious concerns about the current structural behavior and performance of these structures in Romania, especially for those located in strong seismic areas. Additionally, the issue of interventions concerning the possible strengthening of these structures has been considered, particularly the lower segment of the uprights located near the support of the vertical frame. The upgrade of the rack structural system based on evaluation and design checks in most of the cases require strengthening, which must be performed under normal operating conditions, to avoid economic losses caused by suspending the current activity.

Fig. 1.
figure 1

Pallet rack storage system in a typical warehouse in Romania and the dimensions of the vertical upright frame.

Full size image

This study contributes to the level of knowledge by presenting a comparative experimental analysis of two strengthening solutions for the base upright profiles of steel storage pallet racks, under operational conditions.

2 Experimental Study

2.1 Sections and Materials

Table 1 presents the specimens, together with material description, reinforcement cross-section profiles and the connections used to create the reinforced specimens.

Table 1. Specimen and materials description.
Full size table

The material used was established by the supplier of the specimens, who was directly interested in this study, because the specified material is used in the production of the rack structures in Romanian market.

The cross-section of the upright profile is presented in Fig. 2a together with a detail of the stiffening, is labelled “R (reference) 85 × 52 p50” and is perforated only on the front side. The thickness of the steel sheet is 2 mm and the material is S355J2C. The side and front views of the profile with perforations are presented in Fig. 2d.

The cross-sections of the reinforcement profiles (see Figs. 2e and 2f) are made of the same materials with the same thickness. The reinforced specimens and their components before the experimental tests are presented in Fig. 3.

Fig. 2.
figure 2

Cross-section details of the upright profile: (a) basic cross-section without reinforcement, labelled (R) 85 × 52 p50; (b) first reinforcement method, labelled (2L) 2L-85 × 52 p50; (c) second reinforcement method, labelled (C) C-85 × 52 p50; (d) side views V1 and V2 of the profiles; (e) L shaped reinforcement cross-section (method 1); (f) C shaped reinforcement cross-section (method 2).

Full size image

In the case of the second reinforcement method (C), only three specimens were tested instead of five due to technical conditions regarding the provision of reinforcement profiles. Both strengthening methods used M8 bolts to fix the additional profiles internally to the original. The bolts were spaced at 250 mm. In the case of (C) C85 × 52 p50 profile there were 3 bolts on the same section in comparison with the first reinforcement method (2L) 2L85 × 52 p50 which used only 2 bolts per section. This connection method allowed minimal interruption in the activity of the storage rack system. The aim was to assess the viability and effectiveness of these solutions in enhancing the compression resistance capacity of the profiles, thereby ensuring a safer and more reliable racking system. It is important to note that the cross-sections of the reinforced upright profiles used in this study are non-compact, which plays a significant role in their buckling behavior and overall structural performance under axial loads.

Fig. 3.
figure 3

Assembled upright specimens: (a) 3D view of the (2L) 2L85 × 52 p50 specimen; (b) assembled specimens (2L) 2L85 × 52 p50; (c) assembled specimens (C) C85 × 52 p50 profile; (d) 2L and C shape reinforcement profiles; (e) cross-sectional view of the reinforcement profiles.

Full size image

2.2 Methodology

The specimens were tested in a universal testing machine with a manual operating system. In accordance with the testing guidelines [4, 13], thick plates fabricated from 30 mm steel sheet featuring a cut-out for the insertion of a 22 mm diameter steel ball were affixed to the ends of the profiles. Corresponding steel plates, identical in thickness and cutout dimensions, were secured onto the testing machine. The placement of the balls coincided with the center of gravity of the net profile cross-section, being a complete hinge capable of transmitting axial compressive loads. A load cell with a maximal operational capacity of 500 kN was fixed at the top part of the machine. Displacement measurements were conducted using two Linear Variable Differential Transducers (LVDT) arranged diagonally to record axial displacements. The loading force was applied monotonically at a rate of 1.5 kN/s until the point of specimen failure. The force and displacement data were systematically captured via the acquisition system and subsequently analyzed to generate the force-displacement curves.

Figures 4 (a-c) present the experimental tests setup for the: (a) non-reinforced specimen, (b) reinforced specimen according to method 1 using 2L-shaped reinforcement profiles and (c) reinforced specimens according to method 2 using a C-shaped reinforcement profile.

Fig. 4.
figure 4

Experimental tests: (a) non-reinforced specimen setup; (b) specimen setup for method 1 reinforcement with 2L-shaped profiles; (c) specimen setup for method 2 reinforcement with C-shaped reinforcement profile; (d) failure mechanism of the non-reinforced specimen; (e) failure mechanism of the specimens with 2L-shaped reinforcements; (f) failure mechanism of the specimens with C-shaped reinforcement.

Full size image

3 Results and Discussion

The total number of specimens tested was 13, of which: 5 were non-reinforced, 3 were reinforced with method 1, and 5 with method 2. The failure mechanisms of the tested specimens followed expected patterns. As shown in Fig. 4d, the non-reinforced specimens exhibited flexural buckling at mid-span around the weak axis of the cross-section. The strengthening method 1 led to a distortional buckling failure mode (Fig. 4e). In the case of the reinforced specimens with method 2 (Fig. 4f), the additional stiffness provided by the “C” reinforcement profile slightly changed the overall behavior. Based on the deformation patterns, the failure mode appears to be distortional buckling too, with a shorter wavelength. This failure is characterized by flange deformations occurring along the upright. In the picture of Fig. 4f, we can see that the column produced a three-wave pattern along its length, which indicates a distortional buckling with shorter wavelength as in Fig. 4g. This typically happens when the stress exceeds the critical stress for distortional buckling in these elements, causing them to buckle outwards or inwards.

The test results were recorded in the data files for the loads and displacements in a consistent time interval of 0.2 s. These were processed using the Origin software, where the curves for each tested specimen and the average curve for the set of tested specimens were plotted.

A comparative graph is presented in Fig. 5, which features the three average curves, representing the load-displacement curves for the three sets of recordings.

The graphs present three distinct load-displacement curves for the tested specimens. The unreinforced profile labelled (R) 85 × 52 p50, demonstrates a maximum load capacity (Fm) of 134.40 kN. This value serves as the benchmark for evaluating the effectiveness of the reinforcement methods applied to the other specimens.

The first reinforcement method, marked as (2L) 2L85 × 52 p50, shows an increased maximum load capacity to 138.91 kN, a modest improvement over the non-reinforced profile. This suggests that while the 2L reinforcement provides additional strength, the increase is not as significant as might be expected. The results show that the “2L” reinforcement configuration has a limited impact on the overall load-bearing capacity of the profile, because flange rotational stiffness is not improved effectively by the 2L reinforcement, connected only to the flange of the upright profile.

Fig. 5.
figure 5

Comparative load-displacement curves for the reinforced and non-reinforced profiles.

Full size image

In contrast, the second reinforcement method, using a C80 × 40 profile and labelled as (C), exhibits a significantly higher maximum load capacity of 168.25 kN. This is a considerable enhancement compared to the non-reinforced and 2L reinforced profiles, the reinforcement contributing to the rotational stiffness of the flanges, creating connection to the web of the original profile too. The C-shaped reinforcement contributes not only to the increased load carrying capacity, but also appears to improve the axial deformation capacity of the profile, as evidenced by its ability to support higher loads at increased displacements.

The reinforced specimens (2L and C) show improved deformation capacity compared to the non-reinforced specimen (R), the C-shaped reinforcement (C) exhibiting the highest displacement before reaching the maximum load.

Comparing the reinforcing solutions, it proves to be more effective the one which have impact on the critical length. The 2L reinforcement seems to have no impact on the distortional wavelength (critical length for distortional buckling not influenced by the connecting bolts – Fig. 4e). In contrast, C reinforcement proves to be more effective, the distortional wavelength is formed between the fixing bolts position (critical length for distortional buckling reduced to the distance between connecting bolts – Fig. 4f). The non-compact nature of the cross-sections significantly contributed to the observed failure modes, particularly influencing the buckling behavior and load capacity under axial compression.

4 Conclusions

The investigation presented in this study offers alternative solutions to improve seismic performance by reinforcement of existing steel storage pallet racks, a subject of growing importance in the context of growing needs for warehouses located in seismic areas. The experimental results emphasize the effectiveness of employing reinforcement methods to enhance the structural performance of cold-formed steel profiles commonly used in cold-formed steel storage rack systems.

Through meticulous experimental procedures, the study has shown that the second reinforcement method, employing a C-shaped profile, resulted in a notable 25% improvement in compression resistance capacity, thus significantly bolstering the structural integrity of the profiles. The first reinforcement method, while producing a modest 4% increase, still provided valuable reinforcement, indicating that even minor modifications can contribute to the overall stability and safety of such systems, but probably economically unfeasible.

It was observed that reinforcement method can be more effective, when buckling mode or buckling length is influenced: from flexural buckling (non-reinforced profile) to distortional buckling (2L reinforcement) or one wave (2L reinforcement) to three waves (C reinforcement). The increase in load carrying capacity was especially evident in the second reinforcement method.

The implications of these findings are two-fold. First, they validate the potential for retrofitting existing pallet rack structures to meet more demanding operational and safety standards, particularly in seismic zones. Second, they provide a benchmark for the industry, suggesting that the integration of reinforcement profiles into newly designed structures could be a prudent standard practice, contributing to the increased reliability of storage systems.

This study underscores the need for ongoing research and development in the field of refurbishment of steel structures, particularly in the face of evolving industrial needs and environmental challenges. The results of this research offer a promising direction for future studies and practical applications in the design and enhancement of reinforcing steel storage racks, aiming for an optimal balance between structural performance, economic viability, and increased safety.

The limitations of the study consist of the following aspects:

  • Sample size and section diversity: The study used a limited number of samples (13 in total), which may not fully represent the wide range of operational conditions and section configurations found in commercial storage systems. Additionally, only two reinforcement methods were tested, which may not cover all possible reinforcement strategies.

  • Scale of specimens: The experiments were carried out on 1 m lengths of upright profiles. This scale may not accurately reflect the behavior of full-size structures under actual loading conditions, especially in varied environmental contexts.

  • Material variability: The profiles were made from a specific type of steel (S355J2C and S355MC). The results might differ with other steel grades or materials commonly used in different regions or industries.

  • Load application: The loading force was applied monotonically at an approximate rate, which simulates a specific type of loading. Real-world scenarios may involve dynamic or cyclic loading, which could influence the performance of reinforcement methods.

Future research may encompass a broader spectrum of materials to assess the versatility of reinforcement methods in various steel sections, steel grades and alternative materials. It would be beneficial to extend the investigation to dynamic and cyclic load tests to better simulate seismic events and more realistic operational conditions. Furthermore, experimental validation through full-scale structural testing could offer a more in-depth understanding of the performance of the reinforcement methods in real-world scenarios.