Roll service-life optimisation using paper quality measurements for condition indication

Abstract

Paper machine sheet dewatering by roll pressing is critical in the establishment of uniform moisture profiles, as well as other mechanical properties of the paper during manufacture. Roll cover profiles are maintained within specification by removing and regrinding the rolls after fixed periods in service. However, analyses of historical data revealed a high variance in the cover wear rates suggesting the use of a condition-based approach to maintenance. In general excessively uneven wear of the cover profile leads to a reduction in press nip load and an increase in paper moisture content. Papermaking process dynamics, however, result in the moisture content also being influenced by other factors including headbox profile, press felt condition and stock variables such as furnish and consistency. The interrelationships between these variables are highly complex and are thus ideally suited for solution by Artificial Neural Networks. With this work a Neural Network model was used to learn the relationships between these variables based on measurements made in a production environment. The model was shown to successfully model the papermaking process and the influences of press roll wear on paper moisture content. The methodology developed allows the progression of cover wear to be monitored using the moisture profile of the paper produced until quality limits are reached, and the roll is replaced on condition. Comparisons indicate significant benefits of the condition-based approach over conventional methods. Furthermore, the same method can be applied to other nipped rolls, such as roll-coaters and soft-calendars, provided the network is customised for those applications.