Application of the continuous wavelet transform for analysis of formation and streaks in fibrous web structures

The purpose of this research was to develop a method to characterize the distribution of material in paper and other fibrous webs using a method based on the continuous wavelet transform. While investigating this subject, two major advantages of the method were discovered. The first is the ability of the method to accommodate stochastic, non-stationary data sets by spatial localization of the spectral analysis. The second is its ability to distinguish flocs and lightweight region in the spectral analysis.; The impetus for using continuous wavelet transform to analyze the structure of webs, and especially machine made papers, was the need to relate the final structure of the product to the forming processes. Given that objective, an existing method that separates the static and the stochastic components of the wavelet based energy spectrum of the cross machine profile was enhanced to account for additional characteristics of machine direction streaks. These included streak intermittency, off axis orientation and oscillation of the streak position in the cross machine direction. The method was validated using simulated and measured images of the paper formation (distribution of mass). This permitted the existing problem of separating different types of streaks from the wavelet energy spectra. The potential for application of the wavelet algorithm for online processing of webs was also examined. The zone-variance effect on the quality of separation of various streaks was studied. Descriptive parameters that can simplify and effectively represent the energy spectra were described and demonstrated. Machine direction variability that was not identified in the simultaneous space-scale analysis can now be incorporated in the analysis by using the spottiness parameter.; The continuous wavelet transform was also used to delineate between heavy weight (floc) and lightweight zones to provide separate spectra for each. The manner in which these spectra change with floc geometry and density was tested by using simulated and actual formation images. Significant difference in the spectra reflected the manner in which the flocs were formed. This will find significant use for characterizing differences in processes that influence the uniformity of the structure and will allow the identification of process improvements.