1. Introduction
In response to record-breaking global warming that has accelerated since the beginning of the 21st century, the International Maritime Organization (IMO) has continuously strengthened regulations on ship emissions. Consequently, the use of alternative, environmentally friendly fuels–such as liquefied natural gas (LNG), methanol, and ammonia–is expected to increase in marine propulsion systems. In parallel with this transition, the methods for supplying such fuels at ports are diversifying and are generally categorized into three types: Pipe to Ship (PTS), Truck to Ship (TTS), and Ship to Ship (STS).
Each method offers distinct advantages and drawbacks. Among them, the STS method allows for the efficient delivery of large volumes of fuel directly to the receiving vessel, even in offshore areas near ports, through the use of a bunkering vessel. When LNG is supplied via a bunkering ship, the refueling process can be carried out simultaneously with cargo loading and unloading, thereby eliminating the need for dedicated bunkering time. This approach, referred to as LNG bunkering simultaneous operations, was first demonstrated globally in Helsinki in 2019. In Korea, a domestic demonstration was conducted by Korea LNG Bunkering Co., Ltd. at Gwangyang Port in 2023. Although this method enhances operational efficiency and economic viability, it also increases the risk level of the operations, thereby underscoring the need for well-defined safety standards. Presently, approval for LNG bunkering simultaneous operations is based on static stability assessments of mooring lines subjected to current and wind loads. However, dynamic stability under wave conditions also plays a critical role in ensuring the safety of moored vessels during simultaneous operations. Therefore, it is essential to accurately evaluate vessel motions and mooring integrity in response to wave-induced forces.
Historically, ship-to-ship fuel transfer has been widely used in crude oil transportation, and the stability of STS operations has been the subject of extensive research. Much of this prior work has concentrated on tandem or side-by-side offloading between ultra-large floating production and storage units—such as FPSOs and FLNGs—and LNG carriers or shuttle tankers, typically in deepwater environments (e.g., Hong et al., 2002; Zhao et al., 2018; Zhou et al., 2023). In contrast, eco-friendly fuel bunkering is generally intended for conventionally sized vessels and conducted in nearshore environments, particularly within port areas. Although severe wave conditions are uncommon in port settings, the influence of wave reflections from quay walls or nearby moored structures must still be taken into account. Moreover, in port environments, the peak period of the wave spectrum frequently aligns with the natural frequency of piston-mode gap resonance between two side-by-side vessels. Under such resonance conditions, significant vessel motions can be induced. Additionally, large wave amplitudes generated by gap resonance can result in substantial transient drift forces acting on the vessels (e.g., Molin et al., 2009; Li and Zhang, 2016; Zou et al., 2024).
In particular, bunkering vessels are generally smaller than receiving vessels, often possessing approximately half the length and a relatively wider beam due to operational constraints. As these vessels are not designed for rapid, long-distance navigation, they tend to exhibit distinct motion characteristics, particularly depending on their relative position during bunkering operations. These differences necessitate the consideration of factors not typically emphasized in conventional side-by-side offshore operations, including shallow water effects, proximity to quay walls, and the relative size disparity between the vessels. Furthermore, vessels must continue to meet existing safety and environmental criteria applicable to cargo operations during LNG bunkering simultaneous operations.
This study assumes a bunkering simultaneous operation scenario involving the LNG bunkering vessel K LNG Dream, developed by the Korea Research Institute of Ships and Ocean Engineering (KRISO). A numerical analysis was conducted to assess vessel motion and safety under wave conditions in a port environment. Section 2 outlines the numerical models and methods used in the study and analyzes motion responses under various conditions. Section 3 explores the characteristics of mooring loads and vessel motion under different mooring system configurations. The final section summarizes the key findings and discusses directions for future research.
2. Numerical Model and Frequency Domain Analysis
2.1 Numerical Model and Method
To simulate LNG bunkering simultaneous operations, two vessels were selected for analysis. The receiving vessel is an LNG-fueled ship with a length of 91 m, and the bunkering vessel is K LNG Dream, developed by KRISO and equipped with a 500 m3 LNG tank (Fig. 1). The principal dimensions of both vessels are summarized in Table 1.
To analyze the wave loads and motion responses acting on the vessels, the study employed ADFLOW (Choi et al., 2001), a higher-order boundary element method (HOBEM) code based on potential flow theory, developed by KRISO. ADFLOW incorporates additional damping terms on the hull surfaces to suppress the sharp increase in gap flow that may occur between two bodies in multi-body simulations (Zalar et al., 2007; Park et al., 2023). To simulate the quay wall, a box-shaped structure with a uniform width extending to the seabed was discretized and included in the numerical model.
The motion response calculations were performed in the frequency domain under regular wave conditions. The equations of motion for the two-body system are expressed as follows:
Here, mjl is the structural mass, ajl is the added mass, bjl is the wave damping coefficient, bvis,jl is the viscous damping coefficient, cjl is the hydrostatic restoring coefficient, ξl is the motion response, A is the wave amplitude, fj is the wave excitation force, and ω is the wave frequency.
Multiple cases were analyzed to investigate the influence of each structural component. First, the motion response of each vessel was calculated independently. These results were used as a baseline to evaluate the interaction effects between the two vessels, as well as the impact of the quay wall. The port conditions were modeled based on those of Busan Port, assuming a water depth of approximately 20 m and a gap distance of 2 m between the vessels during operations (Fig. 2). The complete modeling configuration for the LNG bunkering simultaneous operation is shown in Fig. 3. In this study, wave heading angles were defined such that head waves corresponded to 180°, while beam waves approaching from the direction of the quay wall were defined as 90°.
2.2 Numerical Results
2.2.1 Motion response of the receiving ship at different water depth
To initially evaluate the fundamental motion performance of a single vessel, the six degrees of freedom (6-DOF) motion responses of the receiving ship were analyzed at different water depths, excluding the quay wall. Fig. 4 presents the response amplitude operators (RAOs) calculated at water depths of 20 m and 200 m to assess the effects of depth. For translational motions which don’t have hydrostatic restoring force–specifically surge, sway, and yaw–the responses were observed to increase continuously under long-wave conditions in shallow water. This trend results from the transition to shallow water wave behavior, wherein wave-induced motions become amplified.
2.2.2 Motion response of the bunkering ship at ship to ship (STS) condition
Subsequently, the 6-DOF motion responses of the bunkering ship were examined under STS conditions, where both vessels are positioned side-by-side without mooring lines. The water depth was maintained at 20 m, and the motion RAOs of the bunkering ship were compared between the STS condition and the standalone case (Fig. 5). A notable amplification in sway, heave, and roll motions was observed near a wave period of 7 s under STS conditions. This amplification is attributed to the wall effect produced by the adjacent receiving ship and to radiated waves generated by the receiving ship’s roll motion. Since radiated waves induced by roll motion are typically considered negligible, the wall effect appeared to be the dominant factor in this case, with additional amplification from wave radiation by the receiving ship. The most significant changes were observed under beam wave conditions, where the wall effect is maximized. In contrast, the receiving ship, being significantly larger, was virtually unaffected by the presence of the smaller bunkering ship. Its motion responses closely matched those in Fig. 4 (standalone case) and were therefore not presented separately.
2.2.3 Motion response of ships at bunkering simultaneous operations (STS to quay wall)
To evaluate the motion responses of both vessels during bunkering simultaneous operations, a case was modeled in which the ships were positioned side-by-side adjacent to a quay wall. Fig. 6 illustrates the 6-DOF motion responses of the receiving ship under these conditions. Given that the receiving ship is largely unaffected by the smaller bunkering vessel, the differences from the standalone case in Fig. 4 can be attributed to the influence of the quay wall. The most notable effects were observed under beam wave conditions, where the quay wall’s effect is most pronounced. In this scenario, sway motion responses reached a plateau rather than continuing to increase under longer wave periods. Roll motion responses exhibited sharper resonance at the peak frequency but declined under longer wave conditions. Heave motion responses displayed a complex behavior, with RAO values approaching 2.0 due to reflected waves, varying slightly above or below this value depending on the wave frequency. In the presence of the quay wall, coupling effects between different motion modes became more significant relative to the standalone case. Notably, heave motion exerted a strong influence on the other motion components. This was further verified by imposing a constraint on the heave motion in the simulation, which resulted in a significant reduction in other motion responses, as shown in Fig. 6.
Fig. 7 presents the corresponding motion responses of the bunkering ship under the same conditions. The bunkering ship is influenced by both the adjacent receiving ship and the quay wall. As previously observed under STS conditions, the proximity of the receiving ship led to strong motion amplification near resonance frequencies. Additionally, the bunkering vessel was further affected by wave reflections and heave coupling effects arising from both the receiving ship and the quay wall. Significant variations were observed in the sway, heave, and roll motions of the bunkering ship. Sway motion exhibited sharp amplification near a 7 s wave period due to the receiving ship’s influence but eventually converged to a constant RAO under longer wave periods because of the quay wall’s effect. Roll motion showed a similar pattern. In both cases, when the heave motion of the receiving ship was constrained, the RAOs of the bunkering ship were significantly reduced. The heave motion response itself demonstrated the most complex behavior: it peaked near a wave period of 7 s under beam wave excitation, decreased under longer waves, and then increased again, converging near an RAO value of 2.0.
3. Mooring Systems and Time Domain Analysis
In Section 2, the motion responses of the vessels were analyzed under wave conditions without considering mooring systems. However, during actual port operations, vessels are securely moored prior to conducting cargo-handling or bunkering tasks. This section investigates vessel responses with mooring systems in place. To accurately compute the loads acting on the mooring lines, it is essential to account for the effects of nonlinear forces. Therefore, a time-domain analysis was performed to capture both the motion responses and the associated nonlinear load effects.
3.1 Time-domain Equation of Motion
For the time-domain analysis, Cummins’ equation based on the time-memory function was adopted (Cummins, 1962). The equation is typically expressed as:
Here, mjl is the mass, ajl is the added mass at infinite frequency, Xj is the motion repsonse, Ljl is the retardation function (time-memory function), cjl is the hydrostatic restoring coefficient, Fwave is the wave excitation force, Fdrift is the drift force, and Fmooring is the mooring restoration force. The vessel motion responses are calculated via time integration based on a quasi-static analysis of the mass–damping–stiffness system described on the left-hand side of the equation.
In port operations, the mooring system generally consists of two main components: fenders, which prevent direct collision between the ship and quay wall, and mooring lines, which suppress dynamic motions and prevent separation from the quay wall. For bunkering operations, an additional mooring system is installed between the bunkering ship and the receiving ship. As with quay mooring, this system also includes fenders and mooring lines to ensure stability between the two vessels.
3.2 Fender System
To prevent collisions between the ship and the quay during berthing, fenders are installed between the hull and the quay wall. Fenders absorb the kinetic energy of the vessel through elastic deformation. The fender system is designed such that the berthing kinetic energy is less than the energy absorption capacity of the fender (PIANC, 2002). According to PIANC (The World Association for Waterborne Transport Infrastructure), the berthing energy (Ef) of a vessel is defined as:
Here, M is the vessel mass, V is the berthing velocity, Ce is the eccentricity coefficient, Cm is the added mass coefficient, Cs is the softness coefficient (default 1.0), and Cc is the contact point coefficient (default 1.0). Recommended berthing velocities vary by ship type and displacement. For this study, a recommended berthing velocity of 0.15 −0.2 m/s was applied for vessels with a displacement below 10,000 t. Using PIANC-specified coefficients and the principal dimensions of each ship, the eccentricity and added mass coefficients were calculated as follows: For the receiving ship: Ce = 0.876, and Cm = 1.61. For the bunkering ship: Ce = 0.776 and Cm = 1.55.
This study assumed that two pneumatic fenders of sufficient size were positioned within the inter-vessel and quay gaps. When vessel motion reduces the gap distance, the fender deforms accordingly, generating a reactive force (Fig. 8). The reactive force is calculated as:
Here, ky is the fender stiffness, ly0 is the original fender diameter, and ly is the deformed fender diameter.
The performance of a fender is commonly indicated by its Guaranteed Energy Absorption (GEA), which is defined as the energy absorbed at 60% deformation. This GEA value also forms the basis for calculating the reactive force of the fender. In this study, the stiffness values of the fenders were determined using the GEA force and the corresponding deformation rate. The selected fender model featured a diameter of 1 m, which is appropriate for maintaining a gap within 2 m. According to typical manufacturer data, pneumatic fenders generate reactive forces in the range of approximately 180–300 kN at GEA deformation, depending on the fender length. Accordingly, the stiffness was set to 500 kN/m between the receiving ship and the quay wall, and to 350 kN/m between the receiving ship and the bunkering ship.
3.3 Wave Drift Force and Motion Response with Fender System
To analyze the impact of waves on the mooring system, the motion-induced responses of the fenders under regular wave conditions were examined. First, wave drift forces were calculated to determine the steady forces acting on the ships. Using the pressure integration method, the wave drift force can be expressed as:
Here, ζ is the wave elevation, δ is the displacement vector, nH is the horizontal normal vector, F(1) is the linear wave load, and ξR is the rotational motion. * indicates a complex conjugate, Re{} is a real part, and Wp is the waterplane area.
Using Eq. (5), the frequency-dependent wave drift forces acting on each vessel under beam wave conditions are presented in Fig. 9. The left figure shows the drift force on the receiving ship, which peaks at a wave period of approximately 6 s. In contrast, the bunkering ship experiences a maximum negative drift force at 6 s, which then transitions to a positive peak around 4.8 s. A negative drift force indicates that the force acts in the opposite direction of wave propagation, meaning the bunkering ship is pushed away from the direction of the waves, and thus opposite to the receiving ship.
This study assumed that two pneumatic fenders were installed in parallel between the quay wall and the receiving ship, and an additional two were installed between the receiving and bunkering ships (Fig. 10). Three wave periods under beam wave conditions were simulated using regular waves, and vessel motion responses along with fender reaction forces were examined through time-domain analysis.
Figs. 11–13 illustrate the relative motions between the ships and the corresponding fender reaction forces for each wave period. Fig. 11 represents long waves with a period of 10.5 s. In this case, both ships experience minimal drift forces. As a result, the relative motion does not reduce the gap sufficiently to trigger contact, and no fender reaction force is generated–even in the absence of mooring lines.
Fig. 12 shows the results for a wave period of 6.28 s. At this frequency, the receiving ship experiences a maximum positive drift force, while the bunkering ship experiences a maximum negative drift force. The bunkering ship is thus pushed in the opposite direction to the receiving ship, preventing the gap between them from narrowing enough to activate the fenders. However, the receiving ship is pushed toward the quay wall with considerable force. Once this force reduces the separation beyond a certain threshold (as indicated by the green box), fender reaction force is generated between the ship and the quay.
Fig. 13 presents the case of a shorter wave period of 4.83 s, during which both vessels experience positive drift forces. As the ships move in the same direction, slight pitching motions also occur, resulting in non-horizontal contact angles with the fender. This leads to more complex motion responses and more frequent fender reactions. Despite the relatively modest magnitude of drift forces at higher frequencies, their combined effect with vessel motions can result in frequent activation of the fenders.
3.3 Mooring Arrangement and Working Safety
Ships berthed at a quay wall maintain their position by connecting mooring lines to bollards. Mooring lines are generally classified into four types based on their function: head lines, breast lines, spring lines, and stern lines (Fig. 14). According to the guidelines of the Oil Companies International Marine Forum (OCIMF), it is recommended that all mooring lines be constructed from the same material. In practical applications, however, achieving complete uniformity across all lines is difficult; therefore, it is advised that mooring lines serving the same function employ the same material and possess consistent mechanical properties. In this study, all mooring lines are assumed to have identical stiffness. Because the vessels considered in this analysis are relatively short, breast lines–which typically run perpendicular to the hull–were omitted.
The material properties of mooring lines vary depending on the size of the vessel. Larger vessels require mooring lines with higher stiffness, whereas smaller ships may use lines with lower stiffness. This variation arises from the size-dependent nature of vessel motions and the corresponding loads imposed on the mooring lines. In this study, because both vessels are categorized as small to medium in size, mooring lines composed of synthetic fiber materials with relatively low stiffness were employed.
The mooring system between the two ships was simplified by placing mooring lines only at the bow and stern of the bunkering ship, which is sufficiently small. These inter-ship mooring lines were modeled with lower stiffness than those used for the receiving ship (Fig. 15).
Each mooring line develops proportional tension according to its elongation from the initial length. Pretension was applied by adjusting the initial line length to ensure a baseline tension under static conditions (Fig. 16). This can be mathematically described as follows:
Here, (x0, y0, z0) and (x, y, z) represent the pre- and post-deformation directional displacements. The length of each mooring line was calculated as
l 0 = x 0 2 + y 0 2 + z 0 2 l = x 2 + y 2 + z 2
One important consideration in the mooring arrangement is the minimization of the vertical angle of the lines between the ship and the quay (OCIMF, 2018). If there is a significant elevation difference between the ends of a mooring line, its effectiveness decreases due to reduced force transmission efficiency. Fig. 17 illustrates mooring configurations under high and low tide conditions, as well as various bollard locations on the quay. During high tide, the freeboard of the ship increases, resulting in steeper vertical angles and diminished mooring effectiveness. Fig. 18 presents the effect of varying freeboard height differences–from 2.4 m to 4.0 m–between the quay and the receiving ship, while maintaining the same pretension. As the vertical angle increases from 5.6° to 14.7°, the motion responses of the ship also increase, indicating a decline in mooring performance. In Fig. 19, the bollard location is shifted laterally inward while maintaining the same freeboard height, thereby reducing the vertical angle to 12.5° and improving mooring effectiveness. The OCIMF guidelines recommend that the vertical angle of mooring lines should not exceed 25°.
Many existing studies on moored ship motions are based on frequency-domain or static analysis (e.g., Lee et al., 2003; Kwak and Moon, 2014; Kwon et al., 2021). Additionally, research on operational safety under port conditions typically incorporates the combined effects of waves, currents, and wind loads (e.g., Cho, 2017; Kim et al., 2016; Kim et al., 2017). This study, however, focuses exclusively on a time-domain analysis of wave-induced motions as an initial investigation into the mooring arrangement.
A numerical free-decay test was conducted to identify the natural period of the coupled sway mode between the two vessels, determined to be approximately 23 s. This test involved displacing the bunkering vessel (located on the seaward side) by 1.0 m and simulating its free response in the absence of external forces. The initial pretension in the mooring lines was set to be less than 200 kN between the quay and the ship and less than 100 kN between the two vessels. Because synthetic fiber lines are used, these values are well below the maximum breaking load (MBL), and the tension variations caused by vessel motions remain small. For the environmental conditions, irregular beam waves were generated using a peak period of 4 s and significant wave heights (Hs) of 0.5 m, 0.75 m, and 1.0 m, representative of the wave climate in Busan Port.
To evaluate the results, we referenced existing motion criteria. The Ministry of Oceans and Fisheries (2024) provides port design guidelines that include two standards for working vessels. Table 2 presents the limiting significant wave heights for cargo handling operations, categorized by vessel size: small ships (<500 GT), ultra-large ships (>50,000 GT), and medium-to-large ships (everything in between). Table 3 outlines the allowable motion response criteria based on ship type and cargo equipment. Most values are based on peak-to-peak motion, except for sway, which uses 0-to-peak. Gas carriers, in particular, are assessed based on the motion tolerance of the loading arm. However, bunkering vessels typically use flexible hoses instead of loading arms (Fig. 20). While gas carrier standards were used as a reference, the motion tolerance for bunkering vessels is expected to be more lenient.
Figs. 21 and 22 plot the sway motion responses of the receiving ship and the bunkering ship, respectively, under varying wave heights. At Hs = 0.5 m and 0.75 m, both vessels exhibit sway motions of less than 0.2 m. At Hs = 1.0 m, the responses increase sharply—up to 0.5 m for the receiving ship and 0.7 m for the bunkering vessel. Consequently, the relative sway motion between the ships increases significantly at Hs = 1.0 m (Fig. 23). As sway motion is highly influenced by drift forces, the response tends to rise sharply beyond a certain wave height. Nevertheless, both vessels satisfy the sway motion criteria for gas carriers listed in Table 3.
Figs. 24 and 25 illustrate the heave motion responses. Although gas carriers have no explicit criteria for heave, the receiving ship’s motion remains within 0.2 m (peak-to-peak). On the other hand, the bunkering ship exhibits heave motions reaching 1.0 m, which may be considered large when compared to general cargo vessel standards. However, as LNG bunkering ships use flexible hoses rather than loading arms in STS operations, a 1.0-m heave motion is unlikely to pose significant operational risk.
Figs. 26 and 27 present the roll motion responses. For the receiving ship, roll amplitudes range from 1.3° to 1.9° (peak-to-peak) across wave heights, approaching but not exceeding the 2.0° limit for gas carriers. In contrast, the bunkering ship exceeds this limit at Hs =0.75 m and above, with responses of 1.9°, 3.5°, and 7.5°, respectively. According to Table 2, the bunkering ship qualifies as a medium-to-large vessel, with a handling wave height limit of 0.5 m, under which both vessels remain within acceptable motion criteria. It should be noted that the roll motion in this study was analyzed under beam wave conditions, with an assumed viscous damping term. To further validate these findings, future studies should include independent verification of roll motion for standalone vessels.
Other motion responses–surge, pitch, and yaw–were significantly smaller and are therefore not presented. Overall, motion amplitudes tended to be larger for the smaller bunkering ship, despite both vessels being subjected to the same environmental conditions. As previously noted, LNG bunkering ships do not utilize rigid loading arms, and roll motion alone is unlikely to impose direct constraints on STS bunkering operations. Currently, port approval standards for moored vessel operations are typically based on single-ship criteria. In contrast, LNG bunkering involves STS configurations where both ships are berthed and operating together. Furthermore, because the bunkering ship is considerably smaller, its motion responses under wave conditions are more pronounced. Therefore, approval standards for STS LNG bunkering operations should be adjusted and relaxed relative to those of conventional gas carriers, and further research is necessary to establish such criteria.
4. Conclusions
To evaluate the stability of LNG bunkering simultaneous operations under wave conditions, both frequency-domain and time-domain motion analyses were conducted. In the frequency-domain analysis, the 6-DOF motion characteristics of the vessels were examined under various environment and vessel configurations, accounting for the effects of water depth, ship-to-ship interaction, and quay wall proximity. The following key findings were obtained:
(1) Under shallow water conditions, horizontal motion responses namely surge, sway, and yaw–increase continuously as wave periods become longer.
(2) Due to the disparity in vessel size, the bunkering ship is significantly affected by the roll motion of the receiving ship, whereas the receiving ship remains largely unaffected by the bunkering vessel.
(3) The presence of the quay wall impacts both vessels, resulting in the convergence of their motion responses to steady-state values under long-period wave conditions.
(4) Wave reflections from the quay wall induce coupling among all motion modes, with heave motion exerting a particularly strong influence on the remaining degrees of freedom.
(5) During LNG bunkering simultaneous operations, the bunkering vessel exhibits complex motion behavior near the resonant frequency of the receiving ship due to the combined influence of ship-to-ship interaction and quay wall reflection. In contrast, the receiving ship is primarily affected by the quay wall.
To account for nonlinear mooring effects, a time-domain analysis was also performed. The wave drift forces and fender responses were initially examined. Under short-period wave conditions, in which both vessels experience positive drift forces, the combination of drift and motion responses in the smaller bunkering vessel led to frequent activations of fender contact forces.
Following the installation of both mooring lines and fenders, vessel motion responses were further examined. By varying the vertical angle between the mooring lines and the quay, it was observed that larger angles resulted in reduced restoring performance and amplified vessel motions.
Finally, vessel motion was analyzed under irregular waves with varying significant wave heights. The receiving ship maintained motion responses within the recommended limits for gas carriers up to Hs = 1.0 m. However, the bunkering ship exhibited larger heave and roll motions, with roll motions reaching up to 7.5°, exceeding the gas carrier operational guidelines. At Hs = 0.5 m, both vessels satisfied all motion criteria, which aligns with the design wave height limit for medium-to-large ships under Korean port design standards. This outcome suggests that the wave height limit is more conservative than the motion-based safety thresholds. Nevertheless, additional studies are necessary to validate the roll damping behavior of the bunkering ship and to consider environmental interactions. Furthermore, existing motion criteria for gas carriers are based on vessels equipped with rigid loading arms, whereas bunkering ships typically employ flexible hoses. Therefore, it is reasonable to propose relaxed motion criteria for bunkering vessels. The same rationale applies to the bunkering of other alternative fuels, such as ammonia and methanol, which also use flexible hose systems. Further research is essential to develop revised constraints for these operations. Ultimately, for LNG bunkering simultaneous operations to be permitted–particularly under ship-to-ship and quay mooring conditions–the safety of mooring systems must be sufficiently ensured.
As part of future work, we plan to investigate the influence of vessel size and principal dimensions on motion responses and to establish operational safety criteria specific to LNG bunkering simultaneous operations.
PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation
Print
