Hostname: page-component-7b7567c8c-wxvwp Total loading time: 0 Render date: 2026-11-16T03:53:28.344Z Has data issue: false hasContentIssue false

On the statistical properties of surface elevation, velocities and accelerations in multi-directional irregular water waves

Published online by Cambridge University Press:  11 January 2021

Mathias Klahn*
Affiliation:
Department of Mechanical Engineering, Technical University Of Denmark, DK-2800 Kgs. Lyngby, Denmark
Per A. Madsen
Affiliation:
Department of Mechanical Engineering, Technical University Of Denmark, DK-2800 Kgs. Lyngby, Denmark
David R. Fuhrman
Affiliation:
Department of Mechanical Engineering, Technical University Of Denmark, DK-2800 Kgs. Lyngby, Denmark
*
†Email address for correspondence: matkla@mek.dtu.dk

Abstract

This paper presents a detailed investigation of the role played by the wave steepness in connection with the statistical properties of the surface elevation and fluid kinematics in irregular, directionally spread, deep-water wave fields initially given by a JONSWAP spectrum. Using ensembles of large wave fields obtained from fully nonlinear simulations, we first consider the statistical properties of the surface elevation. In that connection we determine the probability density functions (PDFs) of the surface and crest elevations for wave fields of relatively small to unprecedentedly large steepness, and compare them with theoretical results from the literature in order to establish the latter's accuracy. We then consider certain statistical aspects of the fluid kinematics found at the surface and of the fluid kinematics accompanying large crests, which to our knowledge marks the first investigation of these properties in the literature. We first determine the PDFs of the horizontal fluid velocities and accelerations as well as the vertical fluid acceleration at the surface. Next, we investigate the joint PDF of the surface elevation and each of the velocities and accelerations at the surface, and use it to determine the surface elevations for which the velocities and accelerations at the surface are large. We then present an analysis of the largest fluid velocities and accelerations found in the vicinity of large crests, and compute the PDFs of these quantities. Finally, we consider the PDFs of the location at which the largest velocities and accelerations occur relative to the crest.

Information

Type
JFM Papers
Copyright
© The Author(s), 2021. Published by Cambridge University Press
Figure 0

Figure 1.(a) The total energy and (b) the steepness as a function of time for different values of the initial steepness $\varepsilon _0$. The legend applies to both figures.

Figure 1

Figure 2.(a) The skewness $\mathcal {S}$ and (b) the kurtosis $\mathcal {K}$ as a function of time. The simulated results are averages of 20 realizations of the initial condition, and the legend applies to both figures.

Figure 2

Figure 3.The PDF of $\zeta$ for (a) $\varepsilon _0 = 0.05$ and (b) $\varepsilon _0 = 0.30$ at times $t = 20T_p$, $t = 50T_p$ and $t = 100T_p$. The legend applies to both figures, which have been computed using data from 20 realizations of the initial condition. The dashed line depicts the PDF of a normal distribution with zero mean and unit variance for reference.

Figure 3

Figure 4.The simulated and theoretical PDFs of the non-dimensional surface elevation $\zeta = \eta / \langle \eta ^{2} \rangle ^{1/2}$ for (a) $\varepsilon _0 = 0.05$, (b) $\varepsilon _0 = 0.10$, (c) $\varepsilon _0 = 0.15$, (d) $\varepsilon _0 = 0.20$, (e) $\varepsilon _0 = 0.25$ and (f) $\varepsilon _0 = 0.30$ all at time $t = 50T_p$. Each histogram is constructed from 20 realizations of the initial condition. The legend applies to all figures.

Figure 4

Table 1.The probability of $\zeta$ being larger than 2, 3, 4 and 5, respectively, obtained from simulations and the first-, second- and third-order PDFs from (4.1) for $\varepsilon _0 =$ 0.05, 0.10 and 0.15. For the simulated results of the form $a \pm b$, $b$ denotes the half-width of the 95 % confidence interval. In most cases the statistical uncertainty is smaller than the error introduced when rounding the probabilities to two significant digits, and confidence intervals have therefore been left out. The simulated results are obtained at time $t = 50T_p$ from 20 realizations of the initial condition.

Figure 5

Table 2.Same as table 1 but with $\varepsilon _0 =$ 0.20, 0.25 and 0.30.

Figure 6

Figure 5.The simulated and theoretical PDFs of the non-dimensional crest height $\zeta _c = \eta _c/ \left(\overline {\eta ^{2}}\right)^{1/2}$ for (a) $\varepsilon _0 = 0.05$, (b) $\varepsilon _0 = 0.10$, (c) $\varepsilon _0 = 0.15$, (d) $\varepsilon _0 = 0.20$, (e) $\varepsilon _0 = 0.25$ and (f) $\varepsilon _0 = 0.30$. Each histogram is constructed from 20 realizations of the initial condition and uses data from the time interval $20 \le t \le 100T_p$. The legend applies to all figures.

Figure 7

Figure 6.The average surface elevation in the domain $\{ (x,y) \, | \, |x - x_c| \le \lambda _p \ \text {and}\ |y-y_c| \le \lambda _p \}$ around crests higher than $H_{m0}$ for (a) $\varepsilon = 0.05$ and (b) $\varepsilon = 0.30$ seen from behind (i.e. for $x < x_c$). The vertical scale on the figures is the same. The figures are computed using data from 20 realizations of the initial condition at times later than $20 T_p$.

Figure 8

Figure 7.(a,b) Contour lines of the average surface elevation in the domain $\{ (x,y) \, | \, |x - x_c| \le \lambda _p \ \text {and}\ |y-y_c| \le \lambda _p \}$ around crests higher than $H_{m0}$ for (a) $\varepsilon = 0.05$ and (b) $\varepsilon = 0.30$. The crest elevation has been normalized to 1. (c,d) Comparison of the average surface elevation around crests higher than $H_{m0}$ and the autocorrelation function (4.8) computed at $t = 50T_p$ for (c) $\varepsilon _0 = 0.05$ and (d) $\varepsilon _0 = 0.30$. The legend applies to both (c) and (d). The figures are computed using data from 20 realizations of the initial condition at times later than $20 T_p$.

Figure 9

Figure 8.The PDFs of (a,b) $v_h^{(s)}/v_0$, (c,d) $a_h^{(s)}/a_0$ and (e,f) $a_z^{(s)}/a_0$ at time $t = 50T_p$ for different values of $\varepsilon _0$. The PDFs are computed from 20 realizations of the initial condition. The legends of the first row apply to all rows. The dashed lines depict the first-order results given by (5.5), (5.6) and (5.7), respectively.

Figure 10

Table 3.The modes, means and standard deviations of $v_h^{(s)}/v_0$, $a_h^{(s)}/a_0$ and $a_z^{(s)}/a_0$ at time $t = 50T_p$ as a function of $\varepsilon _0$. The values are computed from 20 realizations of the initial condition.

Figure 11

Figure 9.The joint PDFs of $\eta /H_{m0}$ and (a,d) $v_h^{(s)}/v_0$, (b,e) $a_h^{(s)}/a_0$ and (c,f) $| a_z^{(s)} |/a_0$ at time $t = 50T_p$ for $\varepsilon _0 = 0.05$ (a–c) and $\varepsilon _0 = 0.30$ (d–f). All figures are produced using 20 realizations of the initial condition.

Figure 12

Figure 10.The conditional PDFs of (a,b) $v_h^{(m)}/v_0$, (c,d) $a_h^{(m)}/a_0$ and (e,f) $a_z^{(m)}/a_0$ given that $\eta _c \ge H_{m0}$ for different values of $\varepsilon$. The PDFs are computed from 20 realizations of the initial condition at times later than $20 T_p$. The legends of the first row apply to all rows.

Figure 13

Table 4.The conditional modes, means and standard deviations of $v_h^{(m)}/v_0$, $a_h^{(m)}/a_0$ and $a_z^{(m)}/a_0$ given that $\eta _c \ge H_{m0}$ as a function of $\varepsilon _0$.

Figure 14

Figure 11.The non-dimensional average extremum velocity and accelerations $V_h/v_0$, $A_h/a_0$ and $A_z/a_0$ (full lines) with 95 % confidence intervals (dashed-dotted lines) as a function of the crest height. The averages have been computed in steps of $0.02H_{m0}$ with $\delta = 0.01H_{m0}$, and the legend applies to all figures.

Figure 15

Figure 12.The joint conditional PDF of $(x_{a_h}^{(m)}, y_{a_h}^{(m)})$ given that $\eta _c \ge H_{m0}$ for (a) $\varepsilon _0 = 0.05$, (b) $\varepsilon _0 = 0.10$, (c) $\varepsilon _0 = 0.15$, (d) $\varepsilon _0 = 0.20$, (e) $\varepsilon _0 = 0.25$ and (f) $\varepsilon _0 = 0.30$.

Figure 16

Figure 13.The conditional PDF of $z_{a_h}^{(m)}$ given that $\eta _c \ge H_{m0}$ for different values of $\varepsilon _0$.

Figure 17

Figure 14.The joint conditional PDFs of (a) $( x_{v_h}^{(m)}, y_{v_h}^{(m)})$ and (b) $( x_{a_z}^{(m)}, y_{a_z}^{(m)})$ given that $\eta _c \ge H_{m0}$ for $\varepsilon _0 = 0.15$.

Figure 18

Figure 15.The conditional PDFs of (a) $z_{v_h}^{(m)}$ and (b) $z_{a_z}^{(m)}$ given that $\eta _c \ge H_{m0}$ for $\varepsilon _0 = 0.15$.

Figure 19

Figure 16.The conditional PDF of $a_h^{(c)}/a_h^{(m)}$ given that $\eta _c \ge H_{m0}$ for different values of $\varepsilon _0$.