Vibroacoustic Simulation. Alexander Peiffer

Чтение книги онлайн.

Читать онлайн книгу Vibroacoustic Simulation - Alexander Peiffer страница 46

Vibroacoustic Simulation - Alexander Peiffer

Скачать книгу

      The Bessel functions can be approximated by a series in 2kR taking the first series term of both functions (Jacobsen, 2011)

       bold-italic upper Z equals rho 0 c 0 pi upper R squared left-parenthesis one-half left-parenthesis 2 k upper R right-parenthesis squared plus j StartFraction 8 Over 3 pi EndFraction k upper R right-parenthesis (2.154)

      This expression is valid for ka<0.5. From the imaginary part we get for the mass

       m equals StartFraction upper I m left-parenthesis bold-italic upper Z right-parenthesis Over omega EndFraction equals StartFraction 8 upper R cubed rho 0 Over 3 EndFraction (2.155)

      Assuming a cylindrical volume V=πR2lc of the fluid above the piston we can calculate the length of the moving mass cylinder to be

       l Subscript c Baseline equals StartFraction 8 upper R Over 3 pi EndFraction almost-equals 0.85 upper R (2.156)

      meaning that at low frequencies the piston is moving a fluid layer of 0.85 times the radius acting as an inertia without radiation.

      2.7.4 Power Radiation

      For the radiated power calculation of the piston we took the pressure at the piston surface and integrated the pressure–velocity product over the surface. Due to the fact that the velocity is constant the surface integral involves mainly the pressure as a space-dependent property. In case of vibrating structures with complex shapes of vibration the velocity distribution over the surface is not homogeneous, and we need a more detailed approach.

       bold-italic p left-parenthesis bold r right-parenthesis equals double-integral Subscript negative normal infinity Superscript normal infinity Baseline StartFraction j omega rho 0 Over 2 pi l EndFraction e Superscript minus j k l Baseline bold-italic v Subscript z Baseline left-parenthesis bold r 0 right-parenthesis d bold r 0 with l equals StartAbsoluteValue bold r minus bold r 0 EndAbsoluteValue (2.157)

      In the above equation a function with argument (r−r0) is multiplied by the velocity function for r0 and integrated over the two-dimensional space. Mathematically, this can be interpreted as a two-dimensional convolution in space

       bold-italic p left-parenthesis bold r right-parenthesis equals StartFraction j omega rho 0 Over 2 pi StartAbsoluteValue bold r EndAbsoluteValue EndFraction e Superscript j k StartAbsoluteValue bold r EndAbsoluteValue Baseline asterisk bold-italic v Subscript z Baseline left-parenthesis bold r right-parenthesis (2.158)

       bold-italic p left-parenthesis bold k right-parenthesis equals StartFraction 1 Over 4 pi EndFraction StartFraction rho 0 omega Over StartRoot k Subscript a Superscript 2 Baseline minus k squared EndRoot EndFraction bold-italic v Subscript z Baseline left-parenthesis bold k right-parenthesis with k equals StartAbsoluteValue bold k EndAbsoluteValue (2.159)

      So, we have replaced the expensive convolution operation by a multiplication. This simplification is at the cost of two-dimensional Fourier transforms that are required to get the expressions in wavenumber domain.

      The time averaged intensity of a sound field is given by the product of pressure and velocity (2.45). As the velocity is not uniform over the surface we perform a surface integration over the vibrating area to get the total radiated power

       StartLayout 1st Row 1st Column normal upper Pi 2nd Column equals double-integral Subscript negative normal infinity Superscript normal infinity Baseline one-half upper R e left-parenthesis bold-italic p left-parenthesis bold r right-parenthesis bold-italic v Subscript z Superscript asterisk Baseline left-parenthesis bold r right-parenthesis right-parenthesis d bold r 2nd Row 1st Column Blank 2nd Column equals double-integral Subscript negative normal infinity Superscript normal infinity Baseline double-integral Subscript negative normal infinity Superscript normal infinity Baseline StartFraction j omega rho 0 Over 4 pi l EndFraction upper R e left-parenthesis e Superscript minus j k l Baseline bold-italic v Subscript z Baseline left-parenthesis bold r 0 right-parenthesis bold-italic v Subscript z Superscript asterisk Baseline left-parenthesis bold r right-parenthesis right-parenthesis d bold r 0 d bold r with l equals StartAbsoluteValue ModifyingAbove r With right-arrow minus ModifyingAbove r With right-arrow Subscript 0 Baseline EndAbsoluteValue EndLayout (2.160)

      Thus, for the determination of radiated power a double area integral is required that may become computationally expensive.

      In the above expression we can also switch to the wavenumber domain. In this case the area integration is replaced by an integration over the two-dimensional wavenumber space.

       StartLayout 1st Row 1st Column normal upper Pi 2nd Column equals double-integral Subscript negative normal infinity Superscript normal infinity Baseline one-half upper R e left-parenthesis bold-italic p left-parenthesis bold k right-parenthesis bold-italic v Subscript z Superscript asterisk Baseline left-parenthesis bold k right-parenthesis right-parenthesis d bold k 2nd Row 1st Column Blank 2nd Column equals double-integral Subscript negative normal infinity Superscript normal infinity Baseline one-half upper R e left-parenthesis StartFraction 1 Over 4 pi EndFraction StartFraction rho 0 omega Over StartRoot k Subscript a Superscript 2 Baseline minus k squared EndRoot EndFraction bold-italic v Subscript z Baseline left-parenthesis bold k right-parenthesis bold-italic v Subscript z Superscript asterisk Baseline left-parenthesis bold k right-parenthesis right-parenthesis d bold k 3rd Row 1st Column Blank 2nd Column equals double-integral Subscript negative normal infinity Superscript normal infinity Baseline one-half upper R e left-parenthesis StartFraction 1 Over 4 pi EndFraction StartFraction rho 0 omega Over StartRoot k Subscript a Superscript 2 Baseline minus k squared EndRoot EndFraction v Subscript z Superscript 2 Baseline left-parenthesis bold k right-parenthesis right-parenthesis d bold k EndLayout (2.161)

      The double integral is replaced by a single two-dimensional wavenumber integration. Thus, once the shape function is available the power calculation in wavenumber space is much faster than in real space (Graham, 1996).

      2.7.4.1 Radiation Efficiency

       mathematical left-angle ModifyingAbove v With caret Subscript z Superscript 2 Baseline mathematical right-angle Subscript upper S Baseline equals StartFraction 1 Over upper S EndFraction integral Underscript upper S Endscripts bold-italic v Subscript z Baseline left-parenthesis bold 
						<noindex><p style= Скачать книгу