Ocean Engineering 32 (2005) 1199–1215 www.elsevier.com/locate/oceaneng Effect of shallow and narrow water on added mass of cylinders with various cross-sectional shapes Z.X. Zhou, Edmond Y.M. Lo*, S.K. Tan MPA-NTU Maritime Research Center, School of Civil and Environmental Engineering, Nanyang Technological University, Singapore 639798, Singapore Received 7 September 2003; accepted 7 December 2004 Available online 25 February 2005 Abstract The sway, heave and roll added masses of three uniform cylinders with semi-circular, rectangular and triangular cross-sectional shapes in shallow and narrow water are numerically analysed. The method is based on simulation of the potential flow induced by the cylinder’s mode of motion. The effects of shallow and narrow water on added mass are analysed and presented. It is concluded that the shallow and narrow water effects on added mass depend on the different cross-section shapes of the cylinders. In particular, the water depth effect on sway added mass is stronger than that on heave added mass while the narrow water effect on sway is weaker than that on heave. The shallow water effect on added mass tends to weaken the narrow water effect. Lastly the effect of shallow and narrow water on added mass on a rectangular cylinder is the strongest while that on a triangular cylinder is the weakest. q 2005 Elsevier Ltd. All rights reserved. Keywords: Added mass; Shallow water; Narrow water; Potential flow 1. Introduction Numerous factors affect the motions of a large floating body. The prediction of such motions requires various hydrodynamic coefficients including added mass and damping. When a seabed is close to the floating body, i.e. the water is shallow, the added mass and damping change significantly due to the proximity to the seabed and the more intensive * Corresponding author. Fax: C65 6792 1650. E-mail address: cymlo@ntu.edu.sg (E. Lo). 0029-8018/$ - see front matter q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.oceaneng.2004.12.001 1200 Z.X. Zhou et al. / Ocean Engineering 32 (2005) 1199–1215 free surface fluctuation, e.g. Newman (1985) and Lamb (1932). Barr (1993) reviewed various ship manoeuvring simulation methods, some of which apply to the shallow water case. He ranked those methods in terms of the efficiency of the methods in addressing the shallow water and free surface effects. De Tarso et al. (1996) investigated the added mass and damping of rectangular cylinders mounted at the seabed and presented the shallow water effect on added mass and damping. Aoki (1997) focused on the hydrodynamic coefficients of very large floating structures in shallow water and quantified the depth effect on the hydrodynamic coefficients. Beukelman (1998) deduced the manoeuvring coefficients for a model wing in deep and shallow water from experiments and discussed the changes caused by shallow water. Abul-Azm and Gesraha (2000) proposed an approximation for the hydrodynamics of floating pontoons in shallow water under incident oblique waves and obtained the depth effect on the pontoon hydrodynamics. Lopes and Sarmento (2002) analysed the hydrodynamic coefficients of a submerged pulsating sphere in finite depth using linear wave theory and presented the changes of added mass and damping with water depth. The added mass of a large submerged body is among the hydrodynamic characteristics that play an important role in determining the body motion. Due to large seawater density, the added mass force is often comparable to other terms in the equations of motion and thus cannot be neglected. As such the evaluation of the added mass of a large floating body is of considerable interest in applications such as prediction of seacraft manoeuvrability and control. While results on frequency dependent added mass of floating cylinders with simple shapes in deep-water are available in the literature, the literature on this same topic in shallow water is considerably less. For an object in translatory motion, it is difficult to obtain the zero-frequency added mass based on the pressure integral over the body’s wetted surface. This is because the pressure force as obtained from the potential theory calculation equals zero when the free surface effect is ignored. The double body theory which is based on the assumption that the free water surface fluctuation is weak when combined with conformal mapping techniques, however, proves to be a useful tool to determine the zero-frequency added mass associated with sway, e.g. Newman (1985). The semi-circular cylinder as a simplification of the object’s cross-section is often adopted in the analysis of added mass. The added masses of circular cylinders and cylinders with other cross-sectional shapes in deep water are available, Newman (1985). Different semi-theoretical methods had also been proposed to determine the added mass of circular cylinder in shallow water, e.g. Lockwood-Taylor (1930), and Kennard (1967). Clarke (2001a) used the techniques of conformal mapping and calculated the added mass of circular cylinder in shallow water. He demonstrated the effect of water depth through a comparison between results from different methods based on conformal mapping techniques and concluded that the approach of using a row of distributed dipoles gave the best accuracy. The added mass for the more complex case of elliptical cylinder in shallow water was given by Clarke (2001b) who used a more general mapping technique based on the Schwartz–Christoffel method. Clarke (2003) further used a similar method to calculate the added mass of elliptical cylinder with vertical fin stabiliser in shallow water. Clarke and the others’ work on the zero-frequency added mass does not address the narrow water effect that can produce changes in added mass values similar to those in Z.X. Zhou et al. / Ocean Engineering 32 (2005) 1199–1215 1201 shallow water. Such effects would be important for example in manoeuvring predictions in a narrow waterway such as that investigated by Sarioz and Narli (2003) for manoeuvres of large ships through the Strait of Istanbul. This paper focuses on a general numerical algorithm used to investigate the effects of shallow water and narrow water on the zero-frequency added mass of a cylinder with various cross-sectional shapes. The cross-sections of semi-circle, rectangle and triangle are considered. The solution involves using the finite difference scheme on an unequal mesh to solve the 2D Laplace’s equation that governs the induced flow-field. The cases for sway, heave and roll are simulated. 2. Mathematical model and equations Fig. 1(a)–(c) show the three cylinders: semi-circular, rectangular and triangular crosssections in shallow and narrow water, along with the Cartesian coordinate system used. A flow-field is induced as the cylinder moves or rotates around its center. The flow boundary comprises three flat planes: the seabed and two side-walls (i.e. seashores), two water levels and the surface of the body. The seawater could be in motion as well. The co-ordinate system with the origin set at the center of the object at the water-plane, the y sway axis and z heave axis is used to describe the flow induced by the cylinder’s mode of motion. The seawater is assumed to be inviscid and incompressible, and the induced water motion irrotational. The induced flow velocity potential is governed by the Laplace’s equation as Dfi Z 0 (1) Ð i whose in which fi is the induced velocity potential by the cylinder’s mode of motion U component in the ith direction is unity and zero for the other five components. Here the subscript i varies between 2, 3 and 4 corresponding to sway, heave and roll, respectively. The water-side surfaces of two static side-walls are considered rigid and impermeable, i.e. vfi j Z0 vy jyjZb (2) in which b is the half of the distance between the two side-walls. In general, there is a free surface effect at the water surface but the effect is ignored here because the zero-frequency case is addressed using the twin-hull approximation. Then the sway case is solved by using a zero vertical velocity condition at the free surface leading to vf2 j Z0 vz zZ0 (3) The heave case is different from the sway case in that it requires that the water level be permeable only in the vertical direction. The following boundary condition at the water level, based on the twin-hull approximation, is adopted: f3 jzZ0 Z 0 (4) 1202 Z.X. Zhou et al. / Ocean Engineering 32 (2005) 1199–1215 Fig. 1. (a) Semi-circular cylinder in shallow and narrow water. (b) Rectangular cylinder in shallow and narrow water. (c) Triangular cylinder in shallow and narrow water. For the roll case, the water level must be permeable to the vertical water flow in order to satisfy the global mass conservation law. Thus similar to the heave case, the boundary condition is imposed: f4 jzZ0 Z 0 (5) The cylinder body surface is assumed to be rigid and impermeable such that the following wall condition is applied. Depending on the body mode of motion, this is given by vfi Ð i $Ð j ZU nh vn rÐZÐr h (6) Z.X. Zhou et al. / Ocean Engineering 32 (2005) 1199–1215 1203 where nÐh is the normal unit vector to the cylinder hull surface pointing into the body and rÐh is the position vector that defines the body surface. Three modes Ð 2 Z ð0; 1 m=s; 0; 0; 0; 0Þ), heave of motion are used corresponding to the sway (U Ð Ð (U 3 Z 0; 0; 1 m=s; 0; 0; 0Þ) and roll (U 4 Z ð0; 0; 0; 1 rad=s; 0; 0Þ). Some singular points exist with at least two normal directions in the flow-field models shown in Fig. 1(a)–(c). For example, at any of the four right-angle corners of the flow domain, the following wall conditions apply: vfi vf Z i Z0 vy vz (7) The corners between the water level and cylinder are assumed to have the normal vectors determined by the cylinder body to avoid the singularity. In addition, the normal vector at any bottom corner of the rectangular and triangular cylinders is considered to be the algebraic average of the vectors at the two points adjacent to the corner, thus avoiding the singularity. The seabed is considered rigid and the application of the impermeable wall condition generates vfi j Z0 vz zZKH (8) in which H is the water depth. Based on the spatial distribution of the induced velocity potential derived from the governing equation, Eq. (1), and the associated boundary conditions, Eqs. (2)–(8), the determination of the sway added mass, heave added mass and roll added mass as given by Newman (1985) is maij Z rw #f n i hj ds (9) S where rw is the seawater density and S is the wetted body surface. 3. Numerical method and computational parameters Numerical simulation using finite difference is adopted to generate the velocity potentials induced by the different cylinder modes of motion. A central difference scheme on an unequal mesh is adopted to discretize Eq. (1). The implicit discrete form is given by cdef 1 1 fðj K 1; kÞ C fðj C 1; kÞ fðj; kÞ Z cd C ef cðc C dÞ dðc C dÞ 1 1 C fðj; k K 1Þ C fðj; k C 1Þ (10) eðe C f Þ f ðe C f Þ where the symbols j and k refer to the y-coordinate and z-coordinate of the node (j, k), respectively; c, d, e and f represent the corresponding distances from node (jK1, k) to node 1204 Z.X. Zhou et al. / Ocean Engineering 32 (2005) 1199–1215 (j, k), node (jC1, k) to node (j, k), node (j, kK1) to node (j, k), and node (j, kC1) to node (j, k), respectively. An imaginary layer of nodes is added at every impermeable wall, such as the side-walls and seabed, to define the discrete wall conditions. The discretization of cylinder hull condition is done by means of various first-order difference schemes, to ensure that the cylinder mode of motion is imposed on the hull to induce the flow-field. The backward difference scheme is adopted at the points with negative y-coordinate while the forward difference scheme is applied to the points with positive y-coordinate. These two first-order difference schemes capture the outward disturbance by the rigid cylinder and have enough accuracy to generate added mass values, as is shown below. Eq. (10) constrained by the discrete boundary conditions is solved by iteration. Iteration is carried out until the following convergence criterion is satisfied: M X N X jfc ðj; kÞ K fp ðj; kÞj% 10K8 (11) jZ1 kZ1 where M and N refer to the number of nodes in the y-direction and the z-direction, respectively. In addition, the subscripts c and p mean the current and previous iteration steps, respectively. The cross-section planes of the cylinders (Fig. 1(a)–(c)) are subdivided evenly into sufficient segments to ensure the convergence. More elements are needed for the shallow and narrow water (typically 40). The water depth changes from deep water depth (up to 20r0) to the smallest water depth of 1.1r0 used. The distance between the two side-walls varies over the range of 20r0 to 1.2r0 to capture the narrow water effect. The second-order approximation is adopted in the evaluation of the integral in Eq. (9), leading to the following discrete form: maij Z rw M 1X ½f ðk K 1Þnhj ðk K 1Þ C fi ðkÞnhj ðkÞDlk 2 kZ1 i (12) in which the variable k is the kth segment on the cylinder surface and Dlk represents the linear length of the kth segment. 4. Numerical results and discussion The numerical results on the added masses of the three basic cylinders and the induced flow patterns by modes of motion are presented in Tables 1–3 and Figs. 2(a)–10(c). Table 1 Added mass of semi-circular cylinder in deep water Added mass Sway (ma22/rwpr20) Heave (ma33/rwpr20) Sources Newman’s analytical value (1985) Present simulation (bZ10r0, HZ10r0) Difference (%) 0.5 0.5 0.5026 0.5026 0.52 0.52 Z.X. Zhou et al. / Ocean Engineering 32 (2005) 1199–1215 1205 Table 2 Added mass of rectangular cylinder in deep water Added mass Sway (ma22/rwr20) Heave (ma33/rwr20) Roll (ma44/rwr40) Sources Newman’s analytical value (1985) Present simulation (bZ10r0, HZ10r0) Difference (%) 2.377 2.377 0.3625 2.392 2.392 0.3653 0.63 0.63 0.77 Figs. 3(a)–4(c) show the flow patterns induced by the rectangular cylinder at various modes of motion in the form of velocity magnitude contours normalized with 1 m/s. Figs. 5(a)–10(c) show the added mass elements in sway, heave and roll. Various corresponding results from Newman (1985), Reddy and Arockiasamy (1991) and Clarke (2001a) are also shown in Tables 1–3 and Fig. 2(a) and (b) for comparison. 4.1. Verification of model and algorithm The comparison of the added masses of the three cylinders in deep and wide water (i.e. semi-infinite domain) with the analytical results given by Newman (1985) and computational results listed by Reddy and Arockiasamy (1991) is shown in Tables 1–3. The writers’ calculations are based on a depth HZ10r0 and width bZ10r0. As is shown later, these depth and width values effectively simulate the semi-infinite domain. The comparison indicates that the present numerical method for the semi-infinite domain generates reliable values of added mass for semi-circular, rectangular and triangular cylinders. A further comparison is shown in Fig. 2(a) and (b) for the sway added masses of semicircular and rectangular cylinders as functions of water depth in wide water i.e. side-walls far from the cylinders and width bZ10r0 from the origin. Results from the semitheoretical analysis by Clarke (2001a) and calculations by Reddy and Arockiasamy (1991) are also shown in Fig. 2(a) and (b). All the results show large increase in the added mass with decreasing depth. The plots show that there are only small differences between the writers’ results and those of the references for very shallow water depth. At depth H approaching r0, the added mass values become infinite, a consequence of the assumption that seawater is incompressible and seabed is rigid: cylinders can no longer move in the sway direction. Table 3 Added mass of triangular cylinder in deep water Added mass Sway (ma22/rwr20) Heave (ma33/rwr20) Roll (ma44/rwr40) Sources Reddy and Arockiasamy (1991) Present simulation (bZ10r0, HZ10r0) Difference (%) 1.1938 1.1938 0.10053 1.1981 1.1981 0.10113 0.36 0.36 0.597 1206 Z.X. Zhou et al. / Ocean Engineering 32 (2005) 1199–1215 Fig. 2. (a) Sway added mass of semi-circular cylinder in shallow and wide water (CaZma22/rwpr20 and the subscript d denotes the deep water value). (b) Sway added mass of rectangular cylinder in shallow and wide water (CaZma22/rwr20 and the subscript d denotes the deep water value). 4.2. Flow patterns Fig. 3(a)–(c) present the flow patterns induced by the modes of motion for the rectangular cylinder when the size of flow-domain is 8r0 by 8r0. It can be seen from the contours that the side and bottom walls increase the added mass through the generation of weak wall flows. It is also noted that these effects are weakest for the roll motion. The additional flow generated at the far boundaries is small resulting in only negligible increases in added mass. As such the added mass values approach those of the semiinfinite case. Fig. 4(a)–(c) show the corresponding flow patterns at shallow (HZ1.5r0) and narrow (bZ1.5r0) water. The density of the contour lines increases significantly indicating a significant increase in the induced flow velocity and thus the corresponding increase in added mass. Similar observations are seen for the circular and triangular cylinders, and are not shown. Z.X. Zhou et al. / Ocean Engineering 32 (2005) 1199–1215 1207 Fig. 3. (a) Velocity magnitude contours caused by rectangle sway mode of motion (bZ8r0, HZ8r0 and contour increment is 0.02). (b) Velocity magnitude contours caused by rectangle heave mode of motion (bZ8r0, HZ8r0 and contour increment is 0.0285). (c) Velocity magnitude contours caused by rectangle roll mode of motion (bZ8r0, HZ8r0). 4.3. Added mass in shallow and narrow water Fig. 5(a) shows the depth variation of the sway and heave added masses of the circular cylinder when the side-walls are far at bZ10r0. The variations of the added mass of the circular cylinder in deep water (HZ10r0) caused by varying the distance to side-walls are shown in Fig. 5(b). Both the sway and heave added masses increase as the water depth 1208 Z.X. Zhou et al. / Ocean Engineering 32 (2005) 1199–1215 Fig. 4. (a) Velocity magnitude contours caused by rectangle sway mode of motion (bZ1.5r0, HZ1.5r0 and contour increment is 0.056). (b) Velocity magnitude contours caused by rectangle heave mode of motion (bZ1.5r0, HZ1.5r0 and contour increment is 0.08). (c) Velocity magnitude contours caused by rectangle mode of motion (bZ1.5r0, HZ1.5r0 and contour increment is 0.035). decreases. The sway case depicts a larger increase. The water depth effect is small until a depth H/r0 of about 2.5 or less. The deep water result is obtained when the depth H/r0 increases to eight or the larger. Similar effects are seen in the case of narrow and deep water (Fig. 5(b)). The two added mass elements increase with decreasing width b/r0. However, the increase in heave is more pronounced. The narrow water effect is small until the width b/r0 is about 2.5 or less. The narrow water effect becomes negligible only at distances larger than about 8r0. Z.X. Zhou et al. / Ocean Engineering 32 (2005) 1199–1215 1209 Fig. 5. (a) Added mass of semi-circular cylinder in shallow and wide water (bZ10r0). (b) Added mass of semicircular cylinder in narrow and deep water (HZ10r0). Fig. 6(a) and (b) show the added mass values for the circular cylinder versus water depth H/r0 at different values of side-wall distance b/r0. When the water depth H/r0 is smaller, the added mass increase caused by decreasing side-wall distance b/r0 is smaller, indicating that the shallow water effect weakens the narrow water effect. The roll added mass is, of course, zero for the circular cylinder as the effect of viscosity is neglected. The changes of the sway, heave and roll added masses of the rectangular cylinder in wide water (bZ10r0) at various water depths are shown in Fig. 7(a). Fig. 7(b) shows the added mass of the rectangular cylinder in narrow and deep water (HZ10r0). Fig. 8(a)–(c) present the added mass of the rectangular cylinder as a function of water depth at various values of the side-wall distance. Figs. 7(a)–8(c) depict that the roll added mass increases with decreasing water depth and/or side-wall distance but its relative increase is by far less 1210 Z.X. Zhou et al. / Ocean Engineering 32 (2005) 1199–1215 Fig. 6. (a) Sway added mass of semi-circular cylinder in shallow and narrow water (CaZma22/rwpr20). (b) Heave added mass of semi-circular cylinder in shallow and narrow water (CaZma33/rwpr20). than those of the sway and heave added masses. This indicates that the roll added mass is little changed by the shallow and narrow water. The trends of the shallow water effect and narrow water effect on the rectangular cylinder (Fig. 8(a)–(c)) are similar to those on the circular cylinder. In particular, the added mass increase due to decreasing side-wall distance b/r0 is smaller at smaller water depth H/r0. Figs. 9(a)–10(c) present the added mass values of the triangular cylinder in different depth and width of the fluid domain such as deep and wide water (HZ10r0, bZ10r0), shallow and wide water (bZ10r0), narrow and deep water (HZ10r0) and shallow and narrow water. It is seen from these figures that the added mass trends on the shallow and narrow water effects on the rectangular cylinder also hold for the triangular cylinder. It is also deduced from Figs. 5(a)–10(c) that the added mass of the rectangular cylinder has the largest relative increase in response to the change of water depth and/or Z.X. Zhou et al. / Ocean Engineering 32 (2005) 1199–1215 1211 Fig. 7. (a) Added mass of rectangular cylinder in shallow and wide water (bZ10r0). (b) Added mass of rectangular cylinder in narrow and deep water (HZ10r0). side-wall distance. The triangular cylinder shows the smallest relative increase. This observation implies that the shallow and narrow water effect on the added mass of the rectangular cylinder is the strongest and that on the triangular cylinder the weakest. This is the direct consequence of the fact that the cylinders with the same breadth at water level and draft have different areas: the rectangle area is the largest and the triangle area is the smallest. 5. Conclusion An algorithm for evaluating the added mass of the semi-circular, rectangular and triangular uniform cylinders in shallow and narrow water has been developed. 1212 Z.X. Zhou et al. / Ocean Engineering 32 (2005) 1199–1215 Fig. 8. (a) Sway added mass of rectangular cylinder in shallow and narrow water (CaZma22/rwr20). (b) Heave added mass of rectangular cylinder in shallow and narrow water (CaZma33/rwr20). (c) Roll added mass of rectangular cylinder in shallow and narrow water (CaZma44/rwr40). Z.X. Zhou et al. / Ocean Engineering 32 (2005) 1199–1215 1213 Fig. 9. (a) Added mass of triangular cylinder in shallow and wide water (bZ10r0). (b) Added mass of triangular cylinder in narrow and deep water (HZ10r0). The cylinder cross-sections have the same breadth at the waterline and draft. The modes of motion include sway, heave and roll. The writers’ results compared well with the corresponding results by Clarke (2001a), Newman (1985) and Reddy and Arockiasamy (1991) at various limiting cases. The presence of bottom and side-walls induced stronger flow patterns that produce significant increase in the added mass values from those of the semi-infinite domain. Significant increase in the added mass is obtained for both shallow and narrow water. It is also shown that the shallow water effect on sway added mass of any cylinder is stronger than that on heave. However, the narrow water effect on sway is weaker than that on heave. The shallow water has the same effect in changing the roll added mass of the rectangular and triangular cylinder as the narrow water but both effects are weaker. The shallow water effect weakens the narrow water effect. The shallow and narrow water effect on the rectangular cylinder is the strongest, that on the circular cylinder moderate and weakest on the triangular cylinder. 1214 Z.X. Zhou et al. / Ocean Engineering 32 (2005) 1199–1215 Fig. 10. (a) Sway added mass of triangular cylinder in shallow and narrow water (CaZma22/rwr20). (b) Heave added mass of triangular cylinder in shallow and narrow water (CaZma33/rwr20). (c) Roll added mass of triangular cylinder in shallow and narrow water (CaZma44/rwr40). Z.X. Zhou et al. / Ocean Engineering 32 (2005) 1199–1215 1215 Acknowledgements Support by MPA-NTU Maritime Research Center, Nanyang Technological University, Singapore is gratefully acknowledged. References Abul-Azm, A.G., Gesraha, M.R., 2000. Approximation to the hydrodynamics of floating pontoons under oblique waves. Ocean Engineering 27 (4), 365–384. Aoki, Shin-ichi, 1997. Shallow water effect on hydrodynamic coefficients of very large floating structures. Proceedings of the 1997 7th International Offshore and Polar Engineering Conference, May 25–30 1997, Honolulu, HI, USA 4 (1), 253–260. Barr, R.A., Review and comparison of ship manoeuvring simulation methods, Transactions–society of naval architects and marine engineers, 101, 1993 SNAME Annual Meeting, (Sep 14–19), New York, NY, USA, 609–635 Beukelman, W., 1998. Manoeuvering coefficients for a wing-model in deep and shallow water. Part I—experiments and test results. International Shipbuilding Progress 45 (441), 5–50. Clarke, D., 2001a. Calculation of the added mass of circular cylinders in shallow water. Ocean Engineering. 28 (9), 1265–1294. Clarke, D., 2001b. Calculation of the added mass of elliptical cylinders in shallow water. Ocean Engineering 28 (10), 1361–1381. Clarke, D., 2003. Calculation of the added mass of elliptical cylinders with vertical fins in shallow water. Ocean Engineering 30 (1), 1–22. De Tarso, P., Esperanca, T., Hamilton Sphaier, S., 1996. Added mass and damping on bottom-mounted rectangular cylinders. Journal of the Brazilian Society of Mechanical Sciences 18 (4), 355–373. Kennard, E.H., 1967. Irrotational flow of frictionless fluids, mostly of invariable density, David Taylor Model Basin, USA Report. 2299, 108–112 Lamb, H., 1932. Hydrodynamics, sixth ed. Cambridge University Press, Cambridge, England (Reprinted 1945, Dover, New York). Lockwood-Taylor, J., 1930. Some hydrodynamic inertia coefficients. Philosophical Magazine Series 79 (55), 161–183. Lopes, D.B.S., Sarmento, A.J.N.A., 2002. Hydrodynamic coefficients of a submerged pulsating sphere in finite depth. Ocean Engineering 29 (11), 1391–1398. Newman, J.N., 1985. Marine Hydrodynamics, third ed. MIT Press, Cambridge, Massachusetts and London, England. Reddy, D.V., Arockiasamy, M., 1991. Offshore Structures, vol. 1. Krieger Publishing Company, Malabar, Florida p. 1991. Sarioz, K., Narli, E., 2003. Assessment of manoeuvring performance of large tankers in restricted water waterways: a real-time simulation approach. Ocean Engineering 30 (12), 1535–1551.