1: %\documentclass[showpacs,aps,pre]{revtex4}
2: \documentclass[twocolumn,showpacs,preprintnumbers,aps]{revtex4}
3: %\documentclass[twocolumn,showpacs,preprintnumbers,pre,aps]{revtex4}
4:
5: %\documentstyle[aps,pre,12pt]{revtex}amsmath,amssymb
6: %\textheight 9.5in
7: %\textwidth 6.7in \oddsidemargin -0.25in \topmargin -0.7in
8:
9: %\newcommand{\myheadings}{\renewcommand{\markright}{right_head}}
10: %\pagestyle{myheadings}
11:
12:
13: \newcommand{\BE}{\begin{equation}}
14: \newcommand{\EE}{\end{equation}}
15: \newcommand{\BA}{\begin{eqnarray}}
16: \newcommand{\EA}{\end{eqnarray}}
17:
18:
19: \newcommand{\BW}{\begin{widetext}}
20: \newcommand{\EW}{\end{widetext}}
21:
22: \input epsf.tex
23:
24: %\doublespace
25:
26: \begin{document}
27: %\draft
28: %\preprint{0202}
29:
30: %\twocolumn[\hsize\textwidth\columnwidth\hsize\csname@twocolumnfalse\endcsname
31:
32: \title{Sound generated by rubbing objects}
33: \author{Zhen Ye}\affiliation{Wave Phenomena Lab, Department of Physics, National
34: Central University, Chungli, Taiwan, R. O. China}
35: \date{\today}
36:
37: \begin{abstract}
38:
39: In the present paper, we investigate the properties of the sound
40: generated by rubbing two objects. It is clear that the sound is
41: generated because of the rubbing between the contacting rough
42: surfaces of the objects. A model is presented to account for the
43: role played by the surface roughness. The results indicate that
44: tonal features of the sound can be generated due to the finiteness
45: of the rubbing surfaces. In addition, the analysis shows that with
46: increasing rubbing speed, more and more high frequency tones can
47: be excited and the frequency band gets broader and broader, a
48: feature which appears to agree with our intuition or experience.
49:
50: \end{abstract}
51:
52: \pacs{ 43.30.Ma, 43.30.Nb}
53:
54: \maketitle
55:
56: Rubbing objects is an everyday experience. People who have gone
57: through cold winters may have rubbed hands to get warm. Rubbing
58: also occurs in Nature. The earthquake rupture is one of the most
59: familiar but disastrous ones. A common observation is that rubbing
60: can generate sound or noise. The sound generated by rubbing hands
61: is certainly common to nearly every one. The sound generation by
62: ice floe rubbing in the Arctic ocean contributes significantly to
63: the ocean ambient noise\cite{PT,Xie}, and may also help sea
64: animals in finding appropriate holes or breaking segments in ice
65: to breath. The sound generated by rubbing scoops against frying
66: pans, however, is likely something most people would like to
67: avoid.
68:
69: In the present paper, we wish to consider the sound generated by
70: rubbing two finite objects. A model is proposed for the sound
71: generation due to the roughness of the contacting surfaces of the
72: two objects. The sound field is calculated and is analyzed for its
73: relation to the roughness and the rubbing speed. It is shown that
74: sound field can be expressed in terms of a series of eigen-mode
75: excitations, leading to tonal features. It is suggested that when
76: the rubbing speed is low, only low frequency modes are possibly
77: excited. With increasing speed, more and more high frequency modes
78: can be excited, making the frequency spectrum broader. These
79: features appear to be in accordance with our intuition or daily
80: experience.
81:
82: Consider the problem of two object rubbing. The conceptual layout
83: is shown in Fig.~\ref{fig1}. One of the surfaces is moving with a
84: constant velocity along the positive $y$ axis, while the other is
85: assumed to be at rest. For simplicity, we assume that the two
86: objects are identical; for different sizes, the effective rubbing
87: surface will equal that of the smaller one. The shape of the
88: objects is a cubic with thickness $L$, width $W$ and height $H$
89: along the $x$-, $y$-, and $z$-th axes respectively.
90:
91: \begin{figure}[hbt]
92: \vspace{10pt} \epsfxsize=1.5in\epsffile{fig1.eps}\smallskip
93: \caption{Conceptual layout of the system.} \label{fig1}
94: \end{figure}
95:
96: When rubbing occurs, a shear displacement field will be generated
97: at the surfaces and is revealed as the shear waves. Such shear
98: waves can leak out through defects or radiate out at the
99: boundaries, and transform into the sound we hear or record by
100: machines, e.~g. microphones.
101:
102: The governing equation for the shear waves inside the objects is
103: \BE \left(\frac{\partial^2}{\partial t^2} - c_s^2 \nabla^2\right)
104: u(x,y,z,t) = 0, \label{eq:we}\EE where $c_s$ is the shear speed of
105: the objects.
106:
107: The boundary conditions are \cite{Bre} \BE \left. \frac{\partial
108: u}{\partial z}\right|_{z=0, H} =0, \ \ \left. u\right|_{y=0,W} = 0
109: \label{eq:bc1},\EE and \BE \left. \frac{\partial u}{\partial
110: x}\right|_{x=L} = 0, \ \ \left. \mu\frac{\partial u}{\partial
111: x}\right|_{x=0} = F(y,z,t), \label{eq:bc3} \EE where $\mu$ is the
112: shear modulus of the material.
113:
114: The stress can be separated into two parts: the mean and the
115: variation, {\it i.~e.} $ F(t, y,z) = \langle F(t,y,z)\rangle +
116: \Delta F(y,z,t)$; here $\langle \cdot\rangle$ denotes the ensemble
117: average. We assume that $\langle F(t,y,z)\rangle$ is independent
118: of $y,z$. It can be shown that the averaged stress can only excite
119: the uninteresting zero-mode, and will thus be ignored.
120:
121: By Fourier transformation, \BE u(x,y,z,t) = \int d\omega
122: e^{-i\omega t} \hat{u}(x,y,z,\omega). \label{eq:ft}\EE
123: Eq.~(\ref{eq:we}) becomes \BE (\nabla^2 + k^2_s) \hat{u}(x,y, z,
124: \omega) = 0, \label{eq:we1}\EE with $$k_s = \frac{\omega}{c_s}.$$
125: The general solution to Eq.~(\ref{eq:we1}) with
126: Eqs.~(\ref{eq:bc1}) and (\ref{eq:bc3}) is \BW\BE
127: \hat{u}(x,y,z,\omega) = \sum_{n=0,m=0}^{\infty}
128: A_{mn}\cos(k_{mn}(x-L))\sin(k_my)\cos(k_nz), \label{eq:s1}\EE\EW
129: where
130: $$k_n = \frac{n\pi}{H},\ \ k_m = \frac{m\pi}{W}, \ \ k_{mn} = \sqrt{k_s^2 - k_m^2 -
131: k_n^2},$$ and $m,n$ are positive integers.
132:
133: The coefficient $A_{mn}$ is determined from \BE \sum_{m,n}
134: k_{mn}A_{m,n}\sin(k_{mn}L)\sin(k_my)\cos(k_nz) = \frac{1}{\mu}
135: F(\omega,y,z).\EE Then $A_{mn}$ is solved as
136: \begin{widetext}\BE A_{mn} = \frac{4}{HW[1+\delta_{n,0}][1+
137: \delta_{m,0}]}\frac{1}{k_{mn}\sin(k_{mn}L)} \frac{1}{\mu} \int_0^H
138: dz\int_0^W dy F(\omega,y,z)\sin(k_my)\cos(k_nz), \label{eq:s2}\EE
139: \end{widetext} with $F(\omega,y,z) = \frac{1}{2\pi}\int dt e^{i\omega t}F(t,
140: y,z).$ It is easy to verify that $\langle A_{mn}\rangle$, caused
141: by the averaged stress, is proportional to
142: $\delta_{n,0}\delta_{m,0}.$ Therefore only zero modes are possible
143: for constant stresses. We will not discuss this situation.
144:
145: The purpose to calculate the intensity of the shear field which is
146: related to the relevant sound field. That is, we need calculate
147: the correlation function \BE D(x,y,z,t; x',y',z',t') \ \equiv\
148: \langle u(x,y,z,t)u^\star(x',y',z',t')\rangle. \label{eq:uc}\EE
149: From the above derivation, it is clear that the key is to find the
150: correlation function of the fluctuation part of the stress at the
151: surface, $\langle \Delta F(y,z,t) \Delta F(y',z',t')\rangle $.
152: Obviously the fluctuation is caused by the roughness of the
153: contacting surfaces. If we assume that the roughness is
154: homogeneous and completely random, i.~e. spatially uncorrelated at
155: different point when the system is at rest. When the surface is
156: moving against each other along the $y$-axis, the spatial
157: separation along this direction is correlated at a later time
158: determined as $\Delta/v$. This consideration leads to \BE
159: \langle\Delta F(y,z,t) \Delta F(y',z',t')\rangle = S \delta(z-z')
160: \delta(y-y' - v(t-t')), \label{eq:fc} \EE where $S$ is a strength
161: factor.
162:
163: Applying Eq.~(\ref{eq:fc}), the intensity field can be calculated.
164: The procedure is to substitute Eqs.~(\ref{eq:s1}) and
165: (\ref{eq:s2}) into Eq.~(\ref{eq:ft}), then calculate
166: Eq.~(\ref{eq:uc}) by taking into account of Eq.~(\ref{eq:fc}). We
167: finally get \BW\BA <u(x,y,z,t)u^\star(x,y,z,t)> &=& S\int d\omega
168: \sum_n \cos^2(k_nz) \sum_{m,m'}
169: \cos(k_{mn}(x-L))\sin(k_my)\cos(k_{m'n}(x-L))\sin(k_{m'}y) \times
170: \nonumber\\
171: & & \frac{H}{2}C_{mn,m'n} \frac{k_m}{(\frac{\omega}{v})^2 -
172: k_m^2}\left((-1)^m e^{i\frac{\omega}{v}W} -
173: 1\right)\frac{k_{m'}}{(\frac{\omega}{v})^2 -
174: k_{m'}^2}\left((-1)^{m'} e^{i\frac{\omega}{v}W} - 1\right).\EA\EW
175: This equation can be rewritten as \BW\BE
176: <u(x,y,z,t)u^\star(x,y,z,t)> = \frac{S}{\mu^2}\int d\omega\sum_{n}
177: \cos^2(k_nz)|Q_{n}(x,y,\omega)|^2,\EE where \BE Q_n(x,y,\omega) =
178: \sum_{m} \frac{2}{W\sin(k_{mn}L)} \cos(k_{mn}(x-L))\sin(k_my)
179: \frac{k_m}{(\frac{\omega}{v})^2 - k_m^2}\left((-1)^m
180: e^{i\frac{\omega}{v}W} - 1\right).\label{eq:Q}\EE\EW Therefore the
181: frequency spectrum of the sound intensity field generated by
182: rubbing is \BE {\cal P}(\omega) = \frac{S}{\mu^2}\sum_{n}
183: \cos^2(k_nz)|Q_{n}(x,y,\omega)|^2,\EE with $Q_n$ being given by
184: Eq.~(\ref{eq:Q}).
185:
186: A few general features can be observed from the spectral formula.
187: First, the strength factor $S$ controls the overall sound
188: intensity level. It is expected to depend on a few parameters of
189: the surfaces, such as the friction coefficients and the mechanical
190: properties of the surfaces. Second, due to the factor
191: $\sin(k_{mn}L)$ in the denominator of $Q_n$, the resonance feature
192: is expected to appear when $\sin(k_{mn}L) = 0$, leading to the
193: phenomenon of tonal sounds. This also shows that the thickness
194: mainly defines the resonance feature. Next, the rubbing speed $v$
195: enters $Q_n$ in the form of $\omega/v$ in the denominator. For a
196: fixed frequency, the decreasing of moving speed will reduced the
197: strength of $Q_n$. Therefore with decreasing speed, high frequency
198: components tend to decay accordingly. Furthermore, since $k_{mn} =
199: \sqrt{(\omega/c_s)^2 - (m\pi/W)^2 - (n\pi/H)^2}$, the cut-off
200: frequencies are determined by $(\omega/c_s)^2 - (m\pi/W)^2 -
201: (n\pi/H)^2 \geq 0.$ These features comply with the intuition or
202: experience. It is apparent that the present model can be extended
203: to consider other rubbing surfaces, by adjusting the correlation
204: function of the surface stress. All these features deduced from
205: the theory tend to support the experimental observation of the
206: sound generated by ice-floe rubbing\cite{Xie}. We note that the
207: experimental results have been interpreted by an alternative model
208: \cite{Ye}.
209:
210: \begin{figure}[hbt]
211: \vspace{10pt} \epsfxsize=2.5in\epsffile{fig2a.eps}\smallskip
212: \caption{Frequency spectra for two rubbing speeds.} \label{fig2}
213: \end{figure}
214:
215: To quantify our discussion, let us consider an example. Suppose
216: that $W=H$, $L=0.1W$, and $S$ is a constant. Two moving speeds are
217: considered: $v/c_s = 1/600$ and $1/300$. As an example, we compute
218: the sound field located at $(L,W/2,0)$. The frequency spectra for
219: the two moving speed are presented in Fig.~\ref{fig2}. The units
220: are arbitrary. Here it is clearly shown that there is indeed a
221: cut-off frequency at about 200 for the lower speed curve. Also for
222: the lower speed case, the discussed resonance feature does appear.
223: For the higher moving speed, the spectrum is broader. But the
224: resonance tends to be weaker. And the cut-off frequency is larger
225: (not shown). Compared to the lower speed case, the frequency band
226: is obviously broader. All these support the above qualitative
227: analysis.
228:
229: In summary, a model is established to investigate the sound
230: generated by rubbing two objects. In the model, we have assumed
231: that the contacting surfaces are rough and the roughness is
232: described as completely random. The dependence of the frequency
233: spectrum on the moving speed and sample geometries is discussed.
234: The results seem to be in line with intuition or experience.
235:
236: \begin{thebibliography}{99}
237:
238: \bibitem{PT} I. Dyer, Phys. Today {\bf 41}(1), 5 (1988).
239:
240: \bibitem{Xie} Y. Xie and D. Farmer, J. Acoust. Soc. Am. {\bf 91}, 1423
241: (1992).
242:
243: \bibitem{Bre} L. M. Brekhovskikh, {\it Waves in Layered Media},
244: (Academic, New York, 1980).
245:
246: \bibitem{Ye} Z. Ye, J. Acoust. Soc. Am. {\bf 97}, 2191 (1995).
247:
248: \end{thebibliography}
249:
250: \end{document}
251: