This dissertation has proposed and validated a pioneering approach in modelling the structure of commercial paper. This method, which is called "structure decomposition analysis", is based on the decomposition of a flocculated fibre network into some simple elements such as disks. In this approach, any substructure of a flocculated network (such as a floc) is decomposed into a set of disks with different diameters and appropriate densities. Results of the structure decomposition analysis were well correlated to the observed formation of paper as obtained by video beta radiography of paper samples.

Moreover, a new method to characterize forming hydrodynamics has been prescribed. This technique is based on the reflection of a laser beam from the stock surface. The reflected signal was modelled by a Poisson pulse process. Geometrical features of the stock surface were deduced from the autocorrelation function of the reflected signal. This technique was shown to be sensitive to process variables such as speed, jet to fabric ratio, and slice opening.

Finally, application of these findings in optimizing the formation of paper on commercial papermachines has been illustrated. Quantitative measures of table activity using this new approach were found to relate well to observed formation and, particularly, to formation measurements using the structure decomposition analysis.