In-line monitoring of the application weight of an adhesion promoter and the residual moisture content in textile finishing by hyperspectral imaging using a NIR multiplex spectrometer system

Abstract

The application weight of an adhesion promoter applied to a textile fabric by impregnation as well as the residual moisture content of the material after drying were monitored by hyperspectral imaging. In order to be able to monitor the current state of a complete textile web, which may have a working width of several meters, a multiplex NIR spectrometer system instead of a hyperspectral camera was used. The optical ports of the multiplexer were connected to a few tens of individual fiber-optic sensors, whose front ends were distributed evenly across the working width of the web. The spectrometer unit covered a spectral range from 1100 to 2200 nm. Chemometric techniques, in particular partial least squares (PLS) regression, were used to extract quantitative values of both parameters of interest from the recorded spectral data. The root mean square errors of prediction (RMSEP) determined in an external validation step were found to be 1.4 g/m2 for the application weight and 0.46 wt% for the moisture content, which are roughly the same error limits as those in a previous study using a hyperspectral camera. Consequently, application weight and residual moisture after drying can be predicted in-line with high precision. Moreover, the spatial distribution of both parameters across the textile was investigated. Although the resolution of the multiplex spectrometer system is one order of magnitude lower than that of a NIR hyperspectral camera, inhomogeneities in both application weight and moisture content could be clearly revealed provided that their size was larger than the limit given by the spatial resolution. The results prove that a scalable NIR multiplex spectrometer system instead of a NIR camera can be used for monitoring of the actual state of the finished material across the overall width of the web. Such data may be used for process control in textile converting and similar processes.