Online measurement of bulk, tensile, brightness, and ovendry content of bleached chemithermomechanical pulp using visible and near infrared spectroscopy

Abstract

Current pulp mill practice to determine the pulp properties of bleached chemithermomechanical pulp (BCTMP) is to prepare and test handsheets for properties of interest. While the use of handsheets is well established, tests require well trained technical staff and introduce a significant lag before test results can be acted upon. In this report, a prototype online visible and near infrared (Vis-NIR) sensor that complements handsheet testing was trialed for 16 months at a BCTMP mill. The system consists of a cabinet located near a probe installed on a conveyor belt between a slab press and a bale press. When a bale stops on the conveyor belt, the probe rises and comes into contact with the bale, taking a spectral measurement and predicting its bulk, tensile strength, brightness, and ovendry (o.d.) content. The data presented here covers a period of 100 days of real-time measurements that correlated with laboratory tests with an R2 of 0.88 for bulk, 0.87 for tensile, and 0.92 for brightness. Bulk, tensile strength, and brightness measurements had root mean square error of prediction (RMSEP) values of 0.19 cm3/g, 3.994 Nm/g, and 0.96 brightness points, respectively. Ovendry content measurements trended with the mill’s existing online moisture measurement system and had an R2 of 0.36 and an RMSEP value of 1.04%. The quality of the correlation for o.d. content was substantially impacted because of difficulties with synchronizing the time stamps between the mill’s existing online measurements and the online system reported here. The online Vis-NIR sensor provides information in real time and has the potential to be used to reach target grade properties more quickly, assist in the identification of inefficient chemical dousing, reduce the effect of operator bias during testing, allow for the prioritizing of laboratory tests, and improve segregation of final products according to customer needs. However, further development would be necessary to use these measurements as part of a process control strategy.