Measurement Error Correction for Distributed Properties -Filter Cake Thickness Measurements

摘要

The measurement of a distributed property is frequently encountered in engineering applications.The impact of a possible measurement error on the measured distribution is described and two methods to account for such an error are detailed: A Fourier transform based decon-volution procedure and a moment based deconvolution approach via kernel estimators are presented.The deconvolution methods can be used in many particle characterization and process applications where particle features or properties are commonly presented in a distributed form.In the current instance,these methods are carefully explored by using artificially generated data and applied to experimental data obtained from filter cake thickness measurements.