Soft sensor modeling of moisture content in drying process based on LSSVM

Least Squares Support Vector Machines(LSSVM) regression principle and measure methods of moisture content during wood drying were introduced. Wood moisture content is a key parameter for regulating and controlling wood drying proces. In this paper soft sensor model based on LSSVM was established for the weakness of wood moisture content measurement in drying process, and parameters selection adopted improved exhaust algorithm. The simulation results of Fraxinus mandshurica and Xylosma racemosum showed that the LSSVM soft sensor model had well robustness and generalization ability, and could predict wood moisture content measurement in drying process accurately, which offered an effective approach for measuring the parameters in the complicated and nonlinear process of wood drying.