Crashworthiness Evaluation of Reinforced Concrete Duct for Subsea Cable Protection to Stock Anchor Drop Impact
Article information
Abstract
Subsea cables are critical infrastructure that are susceptible to damage from anchor drops, necessitating robust protection systems. This study evaluated the crashworthiness of a duct-type reinforced concrete protection (DRCP) structure subjected to the impact of a 1.27-ton stock anchor. A numerical simulation model was developed using the Coupled Eulerian–Lagrange (CEL)-based fluid–structure interaction (FSI) method to replicate actual collision behavior involving the structure and seabed soil. This numerical approach was validated by conducting full-scale crashworthiness tests under conditions equivalent to a free fall from a water depth of 14 meters. The experimental results confirmed that the DRCP structure maintained global structural integrity, sustaining only localized concrete damage. A quantitative comparison of the key impact parameters between the simulation and test data, including maximum compressive strain, maximum strain rate, effective collision duration, and collision strain energy density, showed an average error range of 14.6% to 27.4%. These deviations fall within acceptable engineering limits for the concrete dynamics. Therefore, the proposed FSI-based simulation is a reliable and cost-effective tool for predicting the impact response of subsea cable protection structures, contributing to the development of safer design standards.
1. Introduction
Subsea cables are utilized extensively for power and communication connections between land and islands, as well as for intercontinental transmission and communication networks. The global market for subsea cables is expected to grow at an annual rate of over 17%, with increasing demand observed worldwide (Kim et al., 2019a). As critical infrastructure, any damage to subsea cables results in substantial repair costs and significant social and economic losses. The major risk factors directly causing damage to subsea cables include anchor-drop impacts from vessel navigation, collisions with fishing gear during aquaculture activities, dragging incidents caused by anchoring, and seabed disturbances caused by typhoons or seasonal winds (KEPCO, 2006). According to a survey conducted by the International Cable Protection Committee (ICPC), damage incidents involving subsea cables due to anchor drops in UK waters have recently increased by approximately 40% (ICPC, 2009).
Various protection methods have been implemented to prevent damage to subsea cables, with approaches varying according to factors such as water depth, marine environmental conditions, and seabed characteristics. In the case of soft seabed, subsea cables are typically buried at depths of 1 to 3 meters using equipment such as water jet trenchers or flow modifiers (IEC 62895, 2017; CIGRE Technical Brochures, 2009). Protective structures such as reinforced concrete ducts are installed where seabed conditions involve hard rock or consolidated soils, making burial impractical (Kim et al., 2019a). These protective structures must be designed to withstand extreme conditions such as anchor impacts, but numerical simulations or full-scale tests are required to ensure design safety because there are currently no standardized design codes.
Several studies have been conducted to evaluate the collision performance and structural safety of subsea cable protection structures. Woo and Na (2010) analyzed the maximum collision response of a cylindrical-connected subsea cable protection structure under land-based conditions, considering scenarios involving anchor-drop and drag scenarios. Yoon and Na (2013) assessed the structural response to anchor impact and dragging to propose design criteria by analyzing the stress distributions and deformations of the protective structure under various collision conditions. Woo et al. (2014) applied nonlinear finite element analysis using a smooth particle hydrodynamics (SPH) method to evaluate the impact behavior of a rock-berm type protection structure subjected to anchor drops. Ryu et al. (2015) conducted collision performance simulations of burial-type and flexible concrete mattress-type protection structures under land-based conditions, considering stock anchor drops. Ryu et al. (2016) investigated the influence of the impact angles and reinforcement modeling approaches on the damage assessment of flexible concrete mattress structures subjected to anchor-drop impacts. Jang et al. (2018) conducted numerical simulations using a three-dimensional coupled Euler–Lagrange (CEL) fluid-structure interaction (FSI) approach to realistically replicate the underwater drop phenomena of various hazard apparatuses capable of damaging subsea cables, comparing the drop characteristics depending on the apparatus type. Kim et al. (2019b) used a three-dimensional CEL-based FSI method to simulate underwater collision behavior during anchor drop impacts on subsea cable protection structures. They validated the underwater simulation results by comparing them with those obtained under land-based conditions. Song (2022) conducted collision simulations using the FSI approach to evaluate the risk of anchor-pile penetration into seabed-buried subsea cables and proposed safe burial depths based on comparisons with actual test results. Although previous studies have effectively modeled collision phenomena using nonlinear finite element methods, relatively few have validated their numerical results against full-scale crashworthiness test data, highlighting the need for further test verification.
In this study, a numerical analysis method was developed to assess the crashworthiness of a duct-type reinforced concrete protection (DRCP) structure developed to protect subsea cables from stock anchor drop accidents. The numerical results were verified through a comparison with full-scale crashworthiness tests. If the risk posed by stock anchor drop impacts can be evaluated at a reasonable level through numerical methods, it can significantly reduce the cost and time associated with large-scale verification tests and help develop standard design codes to improve safety. The numerical simulation was performed using a nonlinear finite element analysis (FEA) program based on explicit dynamic integration and a CEL-based fluid-structure interaction (FSI) approach. The actual stock anchor drop accident was simulated by modeling the stock anchor and DRCP structure as Lagrangian domains, while modeling the seabed soil on which the DRCP structure rests as an Eulerian domain. The contact interface between the DRCP structure and the soil was defined as a coupling surface. Full-scale crashworthiness tests were conducted to validate the simulation results. The structural damage and dynamic strain resulting from the stock anchor drop impact were evaluated from the tests. The crashworthiness simulation results demonstrated a strong correlation with the test observations, confirming the structural integrity and impact resistance of the DRCP structure under stock-anchor-drop scenarios. The results of this study are expected to serve as foundational data for design safety codes, ensuring the reliability of subsea cable protection structures. This paper describes the crashworthiness simulation technique and numerical results for the DRCP structure under stock anchor drop impact in Chapter 2, and summarizes the full-scale crashworthiness test procedure and the comparison with simulation results in Chapter 3. Finally, the conclusions are presented.
2. Crashworthiness Simulation
2.1 Concept of CEL-based FSI Analysis
In this study, the CEL-based FSI method was used to evaluate the crashworthiness of the DRCP structure numerically under stock anchor drop impact. As shown in Fig. 1, FSI analysis numerically simulated the interaction between fluid and structure by coupling the Eulerian technique, which defines the fluid domain using the finite volume method (FVM), and the Lagrangian technique, which defines the structural domain using the finite element method (FEM), through a coupling surface (Jang et al., 2018; Kim et al., 2019a; Kim et al., 2019b).
The deformation or motion of the Lagrangian elements (structure mesh) is transferred to the coupling surface shared with the Eulerian elements (fluid mesh), as shown in Fig. 1. Conversely, the pressure generated in the Eulerian domain acts as an external load on the Lagrangian domain through this coupling surface. The common boundary surface where the two media meet was set as the coupling surface to reflect the interaction between the fluid and structural media, enabling the exchange of interaction behaviors. The exchange of forces between the Lagrangian and Eulerian domains at the coupling surface was applied using the penalty method. The penalty method applies a penalty force to both elements simultaneously whenever a relative displacement between the Lagrangian and Eulerian elements is calculated during convergence. The penalty force Fi is defined as follows:
where ks is the contact stiffness defined by material properties; d is the penetration depth, and ni is the unit normal vector at the contact point on the Lagrangian element. In this study, the stock anchor and the DRCP structure were considered as Lagrangian elements, whereas the air layer and the soil constituting the seabed were modeled using Eulerian elements. The MSC.Nastran SOL700, a general-purpose nonlinear FEA program based on the explicit time integration method, was used as the numerical analysis program for the crashworthiness simulation through FSI analysis (MSC.Software, 2017).
2.2 Crashworthiness Simulation Model
The southwest sea of Korea, between the Jeonnam coast and Jeju Island, is an area with the highest concentration of subsea cable installations (Ahn & Kim, 2009). In this study, FSI-based numerical simulations were performed to evaluate the crashworthiness of the DRCP structure under the drop impact of a stock anchor, which has a high frequency of accidents in damage reports of subsea cables installed in this region. The specifications of the stock anchor were selected based on the maximum gross tonnage of vessels mainly navigating the southwest sea, considering a 1.27-ton class stock anchor (Kim et al., 2019b). The structural shape of the 1.27-ton stock anchor was standardized in the Korean Industrial Standards (KS) (KS, 2013). The 3D shape of the stock anchor was implemented in accordance with the design specifications defined in KS. The DRCP structure was designed with rebars embedded in the concrete outer structure to protect subsea cables from falling impact objects and to minimize corrosion from seawater and displacement by currents, considering its seabed installation. Fig. 2 shows the design configurations of the DRCP structure and the stock anchor.
The rebars were arranged in a grid within the DRCP structure to reinforce the concrete, which is vulnerable to tension (Fig. 2(a)). The main external dimensions of the DRCP structure were 9.5 m in length, 3.1 m in width, and 1.48 m in height. The stock anchor consisted of an anchor with a width and height of 1.99 m and 2.95 m, respectively, and a stock of 2.76 m, in accordance with the KS design specifications, as shown in Fig. 2(b).
Fig. 3 shows the FEA model of the Lagrangian domain used to evaluate the crashworthiness of the DRCP structure.
The FEA model of the stock anchor, a hazard factor for subsea cables, used four-node solid elements and consisted of 1,405 nodes and 540 elements. A density of 7,850 kg/m3, an elastic modulus of 210 GPa, and a Poisson’s ratio of 0.3 were applied to reflect the material properties of the stock anchor made of general steel. The weight of the FEA model was verified to match the actual weight. A 100 mm element size was applied to the FEA model of the DRCP structure. The concrete and rebar parts were modeled using hexahedral elements and two-node rebar elements, respectively, consisting of 61,622 nodes and 49,470 elements. The element size in the collision zone between the DRCP structure and the stock anchor was refined to 50 mm, as shown in Fig. 3(a). For the rebar elements, the material properties of a density of 7,850 kg/m3, elastic modulus of 210 GPa, Poisson’s ratio of 0.3, yield strength of 300 MPa, tensile strength of 420 MPa, and elongation of 16% were applied. For the concrete material, a density of 2,230 kg/m3 and a shear modulus of 13 GPa were applied. The material model was considered as a Continuous Surface Cap Model (CSCM) with a compressive yield strength of 45 MPa (Schwer & Murray, 1994). The CSCM is configured so that the limit yield surface decreases in the form of a cap under triaxial compression conditions, allowing a consideration of pore compaction behavior in concrete under high pressure. The strain rate was calculated internally using the effective strain rate based on FHWA (Federal Highway Administration) regulations (Schwer & Murray, 1994; Murray, 2007).
The Euler domain comprises the atmospheric air domain and the soil foundation domain within which the DRCP structure is placed. Fig. 4 presents the FSI analysis model, including the Euler domain configured under the same conditions as the full-scale crashworthiness test.
The FSI analysis model, including the Euler domain, was constructed with dimensions of 11.0 m in length, 4.5 m in width, and 5.0 m in height, as shown in Fig. 4. The dimensions of the ground were determined based on the dedicated test space created for the full-scale crashworthiness test. Based on the literature regarding the accuracy and computation time of FSI analysis (Kim et al., 2019a), the element size of the Eulerian model was set to be 25% finer than that of the Lagrangian model. Hexahedral elements were used for the atmospheric domain, consisting of 377,254 nodes and 355,830 elements, with a density of 1.184 kg/m3 and a bulk modulus of 0.101 MPa. The depth of the soil foundation domain was set to 0.5 m. The soil type was considered to be hard sand, the seabed ground where DRCPs are applied to prevent subsea cable damage, which matched the actual test conditions. Hexahedral elements were used for the soil foundation domain, consisting of 41,917 nodes and 39,536 elements. The material properties of the sand were modeled using the Mohr–Coulomb yield model (Abbo & Sloan, 1995). The soil stiffness properties were determined for the soil at the dedicated test site through site-specific soil tests, including standard penetration tests and vibrocorer drilling tests. Based on the actual geological test results, a density of 2,100 kg/m3, bulk modulus of 0.1 GPa, internal friction angle of 40°, and shear modulus of 83.3 MPa were applied. The coupling effect between the fluid-structure media in the FSI analysis model was accounted for by treating the common boundary surfaces of the stock anchor and DRCP structure as coupling surfaces with the Euler domain.
2.3 Crashworthiness Simulation Results
For the crashworthiness simulation of the DRCP structure caused by stock anchor drop impact, the drop velocity of the stock anchor must first be estimated. In the full-scale crashworthiness test, the stock anchor impactor was dropped from a height of 1.5 m in the direction of gravity using a crane. The drop height of 1.5 m considers the water depth of the southwest coastal area of Korea, where protective structures for HVDC subsea cables, such as DRCP, are mainly installed. The literature reports that the water depth at which subsea cable protection structures are frequently damaged by a dropping anchor is approximately 14 m, which corresponds to an equivalent drop height of approximately 1.5 m under land conditions (KEPCO, 2006; Kim et al., 2019a). The motion velocity and distance for the gravitational drop of the stock anchor can be defined by the following theoretical equations (Kim et al., 2019b):
where ρs and ρa are the densities of the stock anchor and the atmospheric domain, respectively; V is the volume of the stock anchor; g is the gravitational acceleration; Cd is the drag coefficient; A is the projected area of the stock anchor, and v(t) is the drop velocity of the stock anchor. The drop distance x(t) of the stock anchor can be defined as
where ms and ma are the mass of the stock anchor and the added mass of the fluid, respectively, and t is the drop duration. The volume and projected area applied, which are geometric characteristics of the stock anchor, were 0.171 m3 and 0.772 m2, respectively. These values were calculated from the FEA model, and the drag coefficient was derived from Eq. (4). The calculated drop velocity of the stock anchor obtained by applying the geometry and material properties of the FEA model to Eqs. (2)–(4) was 4.9 m/s. The calculated drop velocity can be considered a low-speed loading rate corresponding to the initial stage of resistance-free fall, due to the fluid resistance generated on the stock anchor by the 14 m water depth and high fluid density.
The crashworthiness simulation aimed to compare the results with the full-scale crashworthiness test by measuring the strain generated in the vicinity of the impact location when the stock anchor drops at 4.9 m/s in the direction of gravity and impacts the DRCP structure. The stock anchor was modeled as a rigid body, and the DRCP structure was modeled as a deformable body. The friction coefficient between the stock anchor and the DRCP structure during collision was set to 0.4, which is the recommended friction coefficient between steel and concrete in the literature (PCI, 2017). The time step for the numerical simulation was set to 0.2 ms, with 10 iterations per time step, and the analysis was performed for up to 330 ms. The strain measurement locations and the anchor-drop impact location were set to the upper surface of the DRCP structure, as shown in Fig. 5.
The strain measurement locations were set at 150 mm intervals in all four directions from the center of the anchor projected onto the top surface of the DRCP structure before the drop, as shown in Fig. 5. The anchor drop impact location was set to the position where the actual anchor drops and collides with the top surface of the DRCP structure at a distance of 2 m from the upper right end of the structure, as shown in Fig. 5. The shape of the 1.27-ton stock anchor was asymmetric with respect to the central axis (Fig. 2). Therefore, the position where the anchor drops and collides with the top surface of the DRCP structure deviated by 100 mm and 60 mm in the transverse and longitudinal directions, respectively, from the center of the anchor before the drop, as shown in Fig. 5. This exact deviation measured from the full-scale test was updated directly in the FEA to ensure a substantial correlation with the actual structural damage conditions. The reason for considering crashworthiness by reflecting such a deviated impact position is to evaluate the actual structural damage that can occur when hazard objects with asymmetric shapes drop onto the DRCP structure. Figs. 6 and 7 present the strain measurement method from the crashworthiness simulation and the compressive strain history results for each measurement location, respectively. Table 1 lists the maximum strain value, maximum strain rate, and effective collision duration for each measurement location.
The strain measurement location can be an arbitrary position on the element, as shown in Fig. 6. The strain results are calculated from nodes belonging to the element. Therefore, the strain results from the crashworthiness simulation were calculated by interpolating the results of the four nodes closest to the strain measurement location on the element.
The compressive strain history results, summarized in Fig. 7, show the strain over an analysis time of 330 ms at each measurement location. At the initial moment when the stock anchor impacted the top of the DRCP structure, a rapid increase in compressive strain was observed, followed by an irregular decrease in strain. The irregularity of the strain results after the maximum impact increased as the measurement location was closer to the impact position. In the compressive strain history results of the simulation, the measured strain did not converge to a completely flat state after the maximum compressive deformation immediately following the impact, but continued to exhibit small-scale oscillations. This can be attributed to the combined effects of the high heterogeneity of the model itself because the DRCP structure is a concrete material with embedded rebars, and numerical high-frequency components originating from contact conditions (Riedel et al., 2008), rather than the global natural frequency of the DRCP or the stiffness effect of the seabed. In particular, high-frequency components arising from the repetitive correction of contact stiffness at the contact surface immediately after impact are partially superimposed on the strain response at the measurement location because this simulation used explicit time integration and penalty-based contact. Fig. 7 shows the actual vibration components due to boundary reflections of shock waves, which are reflected in the numerical results. Table 1 lists the maximum strain, maximum strain rate, effective collision duration, and collision strain energy density calculated from the crashworthiness simulation for each measurement location.
The maximum compressive strain results were reviewed by applying the maximum value over the entire crashworthiness simulation time (Table 1). The values were relatively high at SG-#3 and SG-#4, while they were lowest at SG-#2. The magnitude of the maximum compressive strain value appeared higher because it was closer to the drop impact position of the stock anchor. The maximum strain rate (∊̇max) and effective collision duration (Tc) were calculated using the following equations (Table 1):
In Eq. (5), ∊̇(ti) is the strain rate for the entire section. It was calculated using the finite difference method with uniform sampling because the crashworthiness simulation was performed with a constant time step (Δt). In Eq. (6)ton is the time when the relationship (a) SG-#1 measurement location (b) SG-#2 measurement location (c) SG-#3 measurement location (d) SG-#4 measurement location ε(t) ≥ 0.1 εmax is first established; toff is the end time when it returns to ε(t) ≤ 0.1 εmax, and the threshold was set to 10%. A higher strain rate means a faster increase in the strength and stiffness of the material in the impact zone, and the probability of brittle fracture increases as ductility disappears. The strain rates of SG-#3 and SG-#4, which are relatively close to the stock anchor impact position, were calculated to be higher than those of SG-#1 and SG-#2 (Table 1), suggesting that the impact position is directly related to localized damage. A shorter effective collision duration means a larger instantaneous impact force. Short collision duration results of 10.0 to 32.2 ms were calculated at all measurement locations. Hence, the DRCP structure can be viewed as one in which localized micro-damage can occur due to a high instantaneous impact force rather than dispersing impact energy during collision. Collision strain energy density is the elastic energy stored per unit volume when a structure is deformed by an impact force, and is an indicator evaluating how much energy the structure can absorb and withstand. Localized damage may occur at the locations of SG-#3 (6.8 times that of SG-#1) and SG-#4 (22.5 times that of SG-#2) because the magnitude of collision strain energy density is directly proportional to localized damage and shock absorption efficiency, considering that the DRCP structure is made of concrete. Furthermore, as observed in the strain history (Fig. 7), the initial strain peak at SG-#4 is significantly higher than that at SG-#3 because of the asymmetric geometry and the offset impact location of the anchor. Subsequently, the strain at SG-#3 exhibits a distinct rebound phenomenon around 240 ms. This is dynamically correlated with the tilting behavior of the anchor towards the SG-#4 direction after the initial impact rebound, which sequentially relieves the compressive load concentrated on the SG-#3 region.
3. Validation Test and Discussion
The method of evaluating the crashworthiness of a DRCP structure under stock anchor drop impact based on numerical analysis allows for the assessment of dynamic structural characteristics from various aspects, as well as the efficient observation of changes in structural performance due to design modifications. Nevertheless, the results of the crashworthiness simulation of the DRCP structure using FEA based on the explicit time integration method reflect high nonlinear characteristics such as instantaneous collision duration due to drop impact, heterogeneous material properties, and the application of FSI analysis. Therefore, verification through actual tests is necessary for design safety. In this study, a full-scale crashworthiness test was conducted to verify the results of the numerical analysis-based crashworthiness simulation.
For the full-scale crashworthiness test, a stock anchor and a DRCP structure were manufactured to the same specifications as the numerical simulation model to demonstrate the crashworthiness of the DRCP structure under stock anchor drop impact. The verification test was conducted at the KEPCO dedicated test site located in Gochang, Jeollabuk-do, Korea. Fig. 8 presents the fabrication of the actual stock anchor and DRCP structure as well as the configuration of the full-scale crashworthiness test.
The stock anchor was manufactured as a 1.27-ton class based on KS design specifications, as shown on the left of Fig. 8(a). The DRCP structure was manufactured with the same dimensions and materials as the numerical analysis model and placed on the ground of the dedicated test site with a flattened surface, as shown on the right of Fig. 8(a). Although the test site ground was graded and flattened before placing the DRCP structure, achieving perfectly uniform contact is physically impossible. The inherent micro-unevenness of the real ground surface can lead to non-idealized, localized gaps beneath the DRCP. This boundary condition mismatch inevitably produces a difference in support stiffness compared to the perfectly flat, homogeneous Eulerian domain in the simulation, which partially contributes to the variations in the dynamic strain distribution observed between the test and FEA results. Considering that the DRCP weighs approximately 35 tons, however, the slight unevenness and micro-gaps at the interface between the DRCP base and the soil are unlikely to affect the estimation of strain results. The overall test set-up is shown schematically on the left of Fig. 8(b). For the drop impact test, the full-scale stock anchor and DRCP structure were set up, and the stock anchor was dropped from a 1.5 m height using a mobile crane. To accurately set the drop height, a leveling staff was placed at the impact position, and a high-speed camera captured the behavior of the stock anchor and DRCP structure during the collision. As shown on the right of Fig. 8(b), strain gauges with Wheatstone bridge circuits were used for the strain measurement, and the strain gauge bonding locations were set to be identical to those in the simulation to enable a comparison with the crashworthiness simulation results. The test data acquired from the strain gauges were stored in a data acquisition (DAQ) system with a channel-frequency-class low-pass filter applied, and the strain results were processed at a sampling rate of 5 kHz, the same as the crashworthiness simulation analysis.
Fig. 9 shows the compressive strain history measured during the full-scale crashworthiness test with the crashworthiness simulation results, and Table 2 compares the maximum strain. The strain results for a measurement time of 330 ms at each measurement location were summarized and compared with the crashworthiness simulation results, as shown in Fig. 9. The test results showed a similar trend to the simulation results but, overall, exhibited slightly lower values than the analysis results. Regarding the variability of strain response between analysis and test, according to existing research results, the strain response curve due to the impact load measured in the tests typically shows a smooth form, like the test results of this study, due to the strong high-frequency attenuation inside concrete (Goldsmith, 1960; Aggelis, 2011) and the filtering effect of strain gauges and DAQ systems. On the other hand, it was reported that explicit-based numerical analysis exhibits a relatively fluctuating response compared to the test results, as shown in the analysis results of this study, because of the high-frequency numerical oscillations resulting from the high heterogeneity of the concrete material model itself and contact conditions (Jones, 1997; Riedel et al., 2008; Li et al., 2015).
At SG-#4, the strain gauge was damaged, rendering the test results unobtainable. Thus, an analysis of the cause was required from the high-speed camera recordings. In the test as well, a rapid increase in compressive strain appeared at the moment the stock anchor impacted the top of the DRCP structure, followed by a gradual decrease in strain. The high-speed camera results are summarized by major time points and compared with the crashworthiness simulation results in Fig. 10. The event at 10 ms shown in Fig. 10(b) corresponds to the time when maximum compressive strain occurred immediately after the stock anchor impacted the DRCP structure. The stock anchor compressed the upper portion of the DRCP structure while maintaining the drop state, without causing structural collapse. The event at 50 ms is the time when compressive strain decreased rapidly, and the stock anchor appeared to move upward from the top surface of the DRCP structure due to impact rebound, as shown in Fig. 10(c). The event at 200 ms is the time when the inflection point of the increase and decrease in compressive strain occurred, and the stock anchor tilted rapidly in the direction of SG-#4 (Fig. 10(d)). This biased behavior of the stock anchor can be attributed to the asymmetric shape of its upper part, which damaged the strain gauge at the SG-#4 position.
Fig. 11 shows the state of the collision zone of the DRCP structure after the stock anchor drop was finished. The damage to the DRCP structure caused by the stock anchor drop impact resulted in localized breakage only at the collision site, as shown in Fig. 11. The design of the DRCP structure related to crashworthiness can be evaluated as safe because phenomena, such as overall structural collapse of the DRCP structure or the expansion of damage from the collision site to other structural areas, were not observed in the full-scale test upon the drop impact of the 1.27-ton stock anchor.
Tables 2 to 5 compare the maximum strain value, maximum strain rate, effective collision duration, and collision strain energy density derived from the crashworthiness simulation analysis and the test, excluding SG-#4, with the error rates.
The error rate for the maximum strain value was lowest at SG-#2 (10.7%) and highest at SG-#3 (18.3%), which is close to the impact position, as listed in Table 2. The calculated average error rate for the maximum strain was 14.6%, and the agreement between the analysis and test results was judged to be appropriate given that this was a crashworthiness simulation of a concrete structure. As summarized in Table 3, the maximum strain rate results of the analysis and test were similar at SG-#2, but the error rate was high at 36.0% at SG-#3. The average error rate for the maximum strain rate was 21.2%, which was estimated to be reasonable given the entire-section data and the filtering effect of the test measurement results. The effective collision duration results of the analysis and test were similar at SG-#2, but the error rate was high at 46.2% at SG-#3, as shown in Table 4. The average error rate for the effective collision duration was 27.4%, which was judged to be at a reasonable level considering the fluctuation characteristics of the effective section and the filtering effect of the test measurement results. The collision strain energy density results of the analysis and test matched at SG-#2, but showed the highest error rate of 30.2% at SG-#3 (Table 5). The relatively high error rates in effective collision duration (e.g., 46.2% at SG-#3 and 35.3% at SG-#1) are highly sensitive to the 10% threshold condition applied in Eq. (6). In explicit FSI analysis, high-frequency numerical oscillations caused by penalty-based contact corrections prolonged the signal decay time compared to the physical test. In addition, localized concrete microcracking near the impact zone (SG-#3) dissipated energy more rapidly in the real structure, whereas the macro-scale CSCM material model slightly underpredicted this localized damping effect, leading to a time lag in strain recovery. The average error rate for the collision strain energy density was 18.8%, and the agreement between the analysis and test results was judged to be appropriate.
The average error between the crashworthiness simulation analysis and the test for the maximum strain value, maximum strain rate, effective collision duration, and collision strain energy density ranged from 14.6% to 27.4%. The level of error was reviewed based on the literature. A previous study reported that in experiments on the dynamic characteristics of concrete materials, unlike static analysis of metal materials, experimental uncertainties of material heterogeneity and strain rate effects exist in the impact analysis of concrete structures (Bischoff & Perry, 1991). According to previous studies, data scatter in concrete impact experiments themselves is typically reported to be 10%–20% (Zukas, 2004), and a deviation of around 15%–25% between the analysis and test regarding the impact of concrete structures is considered a reasonable prediction range (Saatci & Vecchio, 2009). Therefore, the average error range of 14.6%–27.4% reported in this study is judged acceptable in engineering terms, considering that the overall behavior shape matches as shown in Fig. 9.
4. Conclusions
This study assessed the crashworthiness of a DRCP structure subjected to a stock anchor drop impact using a CEL-based FSI analysis and validated through full-scale crashworthiness tests. The study aimed to establish a reliable numerical method for assessing the structural integrity of subsea cable protection systems under extreme impact scenarios. The following conclusions were drawn from the numerical and experimental results:
(1) Development of FSI-based Crashworthiness Simulation: A numerical simulation model based on the explicit dynamic FEM incorporating CEL-FSI techniques was developed. This model accurately simulated the drop behavior of a stock anchor and its subsequent impact on a DRCP structure, accounting for the complex interactions between the structure and seabed soil foundation.
(2) Verification through Full-Scale Testing: Full-scale crashworthiness tests were conducted using a 1.27-ton stock anchor dropped from a height of 1.5 m, which corresponds to the terminal velocity of a free fall from a depth of approximately 14 m in water. The tests showed that while the DRCP structure sustained localized damage, such as concrete crushing and spalling at the impact zone, it maintained its overall structural integrity without catastrophic failure or global collapse. Hence, the proposed DRCP design possesses sufficient crashworthiness to protect subsea cables from typical anchor-drop incidents.
(3) Validation of Numerical Accuracy: The validity of the proposed simulation method was verified by quantitatively comparing its results with the full-scale test data. The comparative analysis focused on the key impact parameters, including maximum compressive strain, maximum strain rate, effective collision duration, and collision strain energy density; the average error rates were 14.6%, 21.2%, 27.4%, and 18.8%, respectively. Although some discrepancies were observed due to the inherent heterogeneity of reinforced concrete and the high-frequency numerical oscillations typical in explicit dynamics, the results fall within the acceptable range for concrete impact studies (typically 15%–25%), showing the reliability of the analysis.
(4) Engineering Implications and Future Work: The study showed that the CEL-based FSI simulation is a robust and cost-effective tool for predicting the impact response of subsea cable protection structures. This methodology can significantly reduce the reliance on expensive, time-consuming full-scale testing during the design phase. The established numerical procedure will contribute to the development of standardized design codes for subsea cable protection. Future research will extend this methodology to assess the crashworthiness against various types of hazard apparatuses and different seabed conditions to enhance the safety and reliability of offshore infrastructure.
Notes
Chang Yong Song serves on the journal publication committee of the Journal of Ocean Engineering and Technology and had no role in the decision to publish this article. No potential conflicts of interest relevant to this article are reported.
This Research was supported by the Research Funds of Mokpo National University in 2023.
