Optimization of acceleration measurements using PIV

Two methods are presented for extracting accelerations from velocity fields measured by a two camera PIV system. One method traces imaginary particles between the velocity fields, while the other employs linear regression for the direct estimation of the velocity gradients, which are combined to produce the total acceleration. The effect of experimental and discretization errors is discussed and analysed and a simple error parameter for the acceleration computations is identified. Also a filtering technique for the removal of short scale noise in the velocity fields is given. The methods are applied to three different applications, namely Stokes waves, a nearly spilling crest in an irregular wave train and a solitary wave. In all applications, errors and scattering in the accelerations are discussed by comparison to analytical or numerical solutions.