Estimation of Laboratory Wave Reflection by a Transfer Function Method

Abstract

A theory of transfer function method for separating two-dimensional wave data obtained in laboratory experiments into incident and reflected waves is presented in this paper. Based on the linear wave assumption, specific transfer functions are derived from mathematical manipulations of the composite wave field, and the corresponding impulse response functions are obtained by implementing the inverse Fourier transform of transfer functions. These response functions are used to perform convolution integrals with time series data measured by fixed wave gauges at different locations in a wave flume and then to separate the incident and reflected waves. Compared with other available methods, the phase difference between two wave signals is considered in the transfer functions. Thus, the separation of waves does not involve the phase calculation and the corresponding error is avoided. The validity of the present method is examined through numerical examples and laboratory experiments of physical models carried out in a wave flume. A comparison of results from physical experiments shows that the present method gives much better estimates of incident and reflected waves than other methods available in the literature.