Evaluation and prediction of wood properties in pulp and paper production : Estimation of wood chemical properties and pulping results using Near Infrared Spectroscopy: A case study in BillerudKorsnäs in Gävle

• Main Author: Daraei, Mahsa
• Format: Others
• Language: English
• Published: Mälardalens högskola, Akademin för ekonomi, samhälle och teknik 2016
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:se:mdh:diva-32153

Near Infrared spectroscopy has been recently applied in predicting the different materials especially chemical properties of raw wood used in the pulp and paper mills. In this research, the main goal was to estimate wood chemical properties in the chemical pulping at BillerudKorsnäs mill in Gävle. In fact, the quantity of moisture and lignin was measured using a standard method in the laboratory. Then, with near infrared spectroscopy (NIR) data from solid wood, a model was developed to predict moisture and lignin content in the wood chips. Moreover, the yield of digesting process as well as pulp Kappa value were analyzed for the samples taken from the mill.Good calibration models were separately created for Acid Insoluble Lignin (AIL), Total Lignin, Pulp Yield, and Kappa number of pulp using Orthogonal Signal Correction (OSC) treated NIR spectra with the R-Square values of 0.92, 0.89, 0.85, and 0.99, respectively. The model developed for moisture content showed low R-Square value of 0.52, which indicates some over or under estimation in prediction and high calibration errors. So, it cannot be reliably used for prediction. Moreover, Acid Soluble Lignin (ASL) calibration results with R-Square value of 0.077 were poorly correlated with the laboratory measured values.Based on the results, the OSC treated NIR spectra from raw solid wood can be used to estimate AIL and total lignin content as well as pulp yield and Kappa number. These models could be applied to the fiber lines 1 and 2 at the mill to control the pulping process efficiently and increase the pulp quality. However, the reliability of them needs to be analyzed before application.