Multivariate Analysis for the Development of a TMP Simulation Model, 2005 Engineering, Pulping & Environmental Conference

In the TMP process, chip characteristics and process control variables greatly influence pulp properties. Relationships among some of these three categories of variables are sophisticated, elusive, and therefore cannot easily be described precisely. Unfortunately, the research on this field is still immature. Higher customer requirements on pulp qualities call for more in-depth studies. In this research, several dozen variables for chip properties, process control variables and pulp qualities were analysed by PLS. The results show that the chip characteristics can significantly affect the pulp quality, while the process variables for both refining stages have different degrees of influence on pulp properties. A predictive simulation model has been built based on this data analysis to eventually realize the strategy of on-line measurement of chip properties, pulp quality prediction, and automatic adjustment of process variables. The simulation results showed that the errors between the model outputs and the mill data were diminished to a satisfactory level.