High concentration refining(HCR)system is the most critical unit with largest energy consumption in chemithermomechanical pulping(CTMP)production.The modeling and control for HCR is regarded as an important way to improve product quality and reduce energy consumption of the whole pulping and papermaking industry.The HCR is a nonlinear time-varying system with strong coupling dynamic characteristics.Its output fiber morphology are not Gaussian distributed and exhibits strong randomness.Thus it is difficult to establish an effective mathematical model aiming to output fiber morphology of the HCR system with conventional methods.This paper established a probability density function(PDF)model for the output fiber morphology distribution of HCR system based on the theory of stochastic distribution and wavelet neural network(WNN).The complexity of model output variables is simplified via weights decoupling based on B-spline approximation,and the accuracy of the model can be also improved with WNNs powerful nonlinear function approximation and adaptive fault tolerance ability.Finally,the data of a industrial HCR system in practical CTMP process were used to test the model.The results showed that the developed model has high accuracy,strong generalization ability,and better practical value.