Optimizing Refiner Operation with Statistical Modelling

A series of pilot scale refining trials has been conducted to study the impact of refining conditions on the energy efficiency of the process and on the handsheet quality of a chemimechanical pulp. Black spruce (Picea mariana) chips, cooked to a yield of 90% and a sulphonate content of 1.4%, were refined in two atmospheric stages. Plate gap and consistency in each stage was controlled according to a central composite design. Statistical models of refiner performance were constructed from these experimental results and a non-linear optimization of process conditions was conducted. Optimization results were verified with plant trials which indicate that increasing the ratio of specific energy applied in the first stage leads to a reduction of approximately 15% in the total energy required. Inversely, this strategy can be used to obtain significant increases in pulp quality for a given energy input. The results also indicate that the largest proportion of energy must be applied to the fibres at high consistency.