Interference effects on farfield ship waves at high Froude numbers

Francis Noblesse,Renchuan Zhu,Liang Hong,Chenliang Zhang,Jiayi He,Yi Zhu

2013-01-01

Abstract:for IWWWFB 2014 Interference effects on farfield ship waves at high Froude numbers Francis Noblesse, Renchuan Zhu, Liang Hong, Chenliang Zhang, Jiayi He, Yi Zhu State Key Laboratory of Ocean Engineering, School of Naval Architecture, Ocean & Civil Engineering, Shanghai Jiao Tong University, Shanghai, China; noblfranc@gmail.com Introduction and summary of main result The Kelvin wake of a ship of length Ls that advances at constant speed Vs along a straight path in calm water is considered. Basic features of this wake have been explained by Kelvin and are well known. Briefly, Kelvin’s farfield analysis shows that no ship waves can exist outside a wedge, with angles ψ ≈ ±19◦28′ from the ship track, that trails the ship, and that two waves — known as transverse and divergent waves — exist at any point within that wedge. Furthermore, it is also well known that the waves created by a ship mostly stem from the ship bow and stern, where the geometry of a ship hull (a streamlined slender body) varies most rapidly, and that interference between the waves created by these two main wave generators, i.e. the ‘bow wave’ and ‘the stern wave’, result in wave amplitudes that vary within the Kelvin wake. These interference effects are well understood, and well known to naval architects. In particular, interference between the transverse waves generated at a ship bow and stern results in oscillations (‘humps’ and ‘hollows’) in the variation of the wave drag with respect to the ship speed, and avoidance of unfavorable interferences (humps of the wave drag) is an important basic element of ship design that is known to every naval architect. Interference between the divergent waves generated at a ship bow and stern can also occur, although only if the Froude number F ≡ Vs/ √ gLs is sufficiently large. Specifically, interference between divergent bow and stern waves can occur in the ‘high Froude number regime’ Fc < F with Fc ≈ 0.6. This interference between divergent bow and stern waves results in dominant waves, i.e. largest waves, along rays at angles ±ψ from the ship track (x < 0, y = 0, z = 0) defined in terms of the Froude number F as tanψ = √ π2F 4/`2−1 2π2F 4/`2−1 with tanψ ≤ 1/2 √ 2 (1) Here, ` = ξ− ξ denotes the distance between the effective origins x = ξ and x = ξ of the bow wave and stern waves; ` and x are nondimensional in terms of the ship length Ls . Observations show that the effective origins x = ξ and x = ξ are located slightly aft of the ship bow or slightly ahead of the ship stern, and the empirical distance ` = ξ− ξ ≈ 0.9 (2) is a reasonable value that is commonly used by naval architects. The relation (1) yields tanψ = 1/8 and ψ ≈ 19◦28′ for F = Fc with Fc = (3/2) 1/4 √ `/π ≈ 0.6 (3) Thus, the largest waves are found along the cusp of the Kelvin wake for F ≈ 0.6 but inside the Kelvin wake for larger Froude numbers. E.g., (1) yields ψ ≈ 8◦12′ for F = 1. Expression (1) yields the approximation tanψ ≈ `/2 πF 2 as F →∞ (4) which shows that ψ = O(1/F ) as F →∞. Comparison with recently-published alternative results The alternative expression tanψ = √ 2πF 2−1 4πF 2−1 with tanψ ≤ 1/2 √ 2 (5) for the ‘wake angle’ ψ is given in [1] and [2] . This expression yields tanψ = 1/8 for F = Fc with Fc = √ 3/ √ 4π ≈ 0.5 (6) Figure 1: Experimental observations (marked as dots) of the ‘wake angle’ ψ reported in [1] and corresponding theoretical predictions given by expressions (5) or (1) obtained in [1,2] or from an analysis of interference effects between the divergent waves created by a ship bow and stern, respectively. The curves associated with expressions (5) or (1), which predict that ψ is O(1/F ) or O(1/F ) as F → ∞, are identified as F or F , respectively. Thus, (5) predicts that the largest waves are found along the cusp of the Kelvin wake for F ≈ 0.5 but inside the Kelvin wake for 0.5 < F, rather than for 0.6 < F in (3). For F = 1, (5) yields ψ ≈ 11◦23′. Furthermore, (5) yields the approximation tanψ ≈ 1/2 √ 2πF as F →∞ (7) Thus, (5) predicts that ψ is O(1/F ) as F →∞, whereas (4) yields ψ = O(1/F ) as F →∞. Expression (1) is a straightforward consequence of the interference between the divergent waves that are generated at a ship bow and stern, as shown further on. The alternative expression (5) is obtained in [1] by invoking a flow model that assumes that a ship hull of length Ls cannot generate waves of length appreciably larger than Ls , although the wavelength 2πV 2 s /g of the longest waves created by a ship is larger than the ship length Ls if the Froude number F is greater than 1/ √ 2π ≈ 0.4 . Expression (5) is also validated in [1] using numerical computations, and in [2] using both numerical computations and a classical mathematical analysis, of the Kelvin wake generated by a Gaussian distribution of pressure at the free surface. However, such a pressure distribution cannot account for interference effects between the wave systems created by a ship bow and stern, although it may be a reasonable model for the flow around a high-speed boat in the planing regime. Discrepancies between the results obtained in [1,2] for a Gaussian pressure distribution (a wave generator with a single peak) at the free surface and given by an analysis of interference effects between the two dominant waves generated by a ship bow and stern (a ‘two-point’ wave generator) are therefore not surprising. Comparison with experimental observations Fig.1 depicts the theoretical predictions of the ‘wake angle’ ψ given by expressions (5) or (1) obtained in [1,2] or from an elementary analysis of interference effects between the divergent waves created by a ship bow and stern, respectively. The curves associated with expressions (5) or (1), which predict that ψ is O(1/F ) or O(1/F ) as F →∞, are identified as F or F , respectively, in Fig.1. The experimental observations reported in [1] are also included in Fig.1, where they are marked as dots. The figure shows that for F < 0.6, experimental observations are within the range 13◦≤ ψ ≤ 21◦ and are mostly located around the Kelvin angle ψ ≈ 19◦28′. Values of ψ that are smaller than the Kelvin angle are expected as a result of interference between the transverse waves generated at a ship bow and stern. The experimental observations within the range 0.6 ≤ F ≤ 1.1 are in excellent agreement with the theoretical predictions (1) given by an analysis of interference effects between the divergent waves created by a ship bow and stern. The well-established basis of the analysis of interference effects that underlie (1) and the excellent agreement between the theoretical predictions given by (1) and the experimental observations reported in [1] and shown in Fig.1 provide strong evidence that the ‘wake angle’ ψ along which the largest waves created by a ship are found is merely the result of interference between the divergent waves generated by two main wave generators, the ship bow and stern, where the geometry of a ship hull varies most rapidly. The three experimental observations in the range 1.3 < F are significantly higher than the ‘waveinterference’ prediction (1) and in good agreement with the ‘Gaussian free-surface pressure-distribution’ prediction (5). This finding is not surprising because interference between bow and stern waves is much less effective for a high-speed boat in the planing regime than for a displacement ship. Indeed, the flow around a boat in the planing regime can reasonably be represented by means of a free-surface distribution of pressure that has a single peak, but a pressure distribution with two separate peaks is required to realistically model the flow about a displacement ship. An elementary analysis can indeed be performed to compare the ‘wake angles’ ψ produced by single-point or two-point wave generators, and this analysis will be reported elsewhere. Elementary analysis of interference between bow and stern waves Expression (1) is now briefly justified (A more complete justification will be given elsewhere). Steady ship waves are a class of dispersive waves governed by the dispersion relation 1 ≤ F k = 1/cosγ ≡ 1+ tanγ with − π ≤ γ ≤ π (8) This dispersion relation defines two distinct dispersion curves, that are symmetric with respect to both the axis β = 0 and the axis α = 0, in the Fourier plane (α, β) ≡ k(cosγ, sinγ). The wavelength λ ≡ 2π/k of the wave generated at the point (8) of a dispersion curve is given by λ = 2πF cosγ ≤ 2πF 2 ≡ λ (9) This expression shows that the longest waves generated by a ship are shorter than the ship length for Froude numbers F < 1/ √ 2π ≈ 0.4 but are longer than the ship length if 0.4 < F . The Kelvin wake is determined from the dispersion function ∆ = F α−k, as shown in e.g. [3], and is given by { −xn yn } = 2nπF 2 |cosγ | { 1+ sinγ sinγ cosγ } with − π 2 < γ < π 2 and n = 1, 2, 3, . . . (10) The branch π/2 < γ < 3π/2 of the two dispersion curves yields the same wave pattern as the branch −π/2 < γ < π/2, which is then only considered in (10). The parametric equations (10) yield xn < 0. This result means that a ship that advances in calm water only generates waves in the region x < 0, i.e. behind the ship. Furthermore, the wave pattern (10) is symmetric about the ship’s path y = 0. Thus, only the positive half 0 ≤ γ < π/2 of the Kelvin wake is now considered. Farfield waves stem from points of the dispersion curves where the phase θ ≡ αx+βy is stationary. The dispersion relation (8) define two points of stationary phase, which generate two distinct systems of waves known as transverse and divergent waves. The waves generated at a point of stationary phase are radiated in a direction, defined by the angle ψ, that is normal to the dispersion curve and colinear with the group velocity vg ≡ (v g , v g ) as shown in e.g. [3]. Thus, one has tanψ ≡ yn −xn = v g −vx g = −∆β ∆α = β/α 2F 2k −1 = sinγ cosγ 1+ sinγ = tanγ 1+ 2tanγ = tanφ 2 + tanφ (11) Here, φ denotes the angle between the ship track and the wave crests. The relations (11) determin