Optimization of paper pulp production using Artificial Neural Networks and Simulated Annealing
Filip Sandkvist · Lund University Publications Student Papers (Lund University) · 2014
The purpose of this paper is to investigate the possibilities of optimizing the production of paper pulp in a Conical Disc (CD) refiner for different production rates, taking into account the variables that are thought to be the most important for the production. These include three separate dilution waters, conical disc gap and flat gap. Another task is to compare the values of these variables after optimization for different production rates in order to see if there is a pattern between these to be found. These two tasks are solved using Artificial Neural Networks (ANN:s) and Simulated Annealing. Data consists of measurements made on a CD70 refiner, and is provided by Jan Hill at QualTech Jan Hill AB. ANN:s seems to be a possible tool for calculating refiner-specific parameters needed to predict the long fiber content and dewatering capability of the pulp. Simulated Annealing was used in an attempt to show that there exists a relationship between variable values that gives maximum freeness for a fixed long fiber content of 45 %.