Multiresolution analysis and estimation of the basis weight signal in papermaking processes

Abstract

The cross-directional (CD) basis weight signal analysis and estimation for paper machines are investigated, and a new approach is proposed by using the wavelet multiresolution analysis theory. Orthogonal wavelet functions are used to represent CD basis weight variations at various resolutions in the discrete wavelet transform domain. By selecting a suitable threshold resolution, the controllable components can be separated from the original signal. Simulation results indicate that the dimension of the approximating signal can be significantly reduced. It is more suitable for CD modeling and control of basis weight variations on paper machines.