Chen W, Holm S
Simula Research Laboratory, PO. Box 134, 1325 Lysaker, Oslo, Norway.
J Acoust Soc Am. 2003 Nov;114(5):2570-4. doi: 10.1121/1.1621392.
Szabo's models of acoustic attenuation [Szabo, J. Acoust. Soc. Am. 96(1), 491-500 (1994)] comply well with the empirical frequency power law involving noninteger and odd-integer exponent coefficients while guaranteeing causality, but nevertheless encounter the troublesome issues of hypersingular improper integral and obscurity in implementing initial conditions. The purpose of this paper is to ease or remove these drawbacks of the Szabo's models via the Caputo fractional derivative concept. The positive time-fractional derivative is also introduced to include the positivity of the attenuation processes.
萨博的声衰减模型[萨博,《美国声学学会杂志》96(1),491 - 500(1994)]很好地符合了涉及非整数和奇整数指数系数的经验频率幂律,同时保证了因果性,但仍然遇到了超奇异反常积分和初始条件实施中模糊性的棘手问题。本文的目的是通过卡普托分数阶导数概念来缓解或消除萨博模型的这些缺点。还引入了正时间分数阶导数以包含衰减过程的正性。