西北林学院学报2024，Vol.39Issue(1)：223-227,255,6.DOI:10.3969/j.issn.1001-7461.2024.01.30
楸木板材干燥含水率时间序列仿真分析与预测
Prediction and Simulation Analysis on Time Serials of Juglans mandshurica Lumber Moisture Content During Conventional Kiln Drying
摘要
Abstract
In this study,the time series of moisture content in Juglans mandshurica lumber during drying process were simulated and analyzed,and the real-time monitoring and prediction of moisture content in wood drying process were studied.Firstly,the change characteristics of the lumber moisture with time were analyzed,and fitted by the third-order polynomial.Then,the artificial neural network was employed to construct the model for simulation and analysis,and the model was used to predict the wood drying moisture content.The results showed that the moisture content of conventional drying wood presented a non-linear downward trend.The calibration function of water content and time constructed by polynomial fitting was y=60.715 5-0.059 5x-0.000 425 6x2+0.000 000 000 649 6x3,and there existed a large er-ror between function fitting simulation and measured value.The maximum error of the simulation results was 2.87％obtained by constructing BP neural network model with the topology structure of 6×8×1.By the trained BP network model to predict water content,the average error was 3.63％,the maximum and minimum errors were 11.15％and 0.04％.The prediction errors of 3 samples in 25 samples were more than 10％,the prediction errors of the remaining 22 samples were less than 9％.Moreover,local large er-rors occurred in the region with large fluctuation of water content.The BP neural network can be used to predict the drying moisture content of J.mandshurica lumber in one step.关键词
木材干燥/含水率/神经网络/仿真分析Key words
wood drying/moisture content/neural network/simulation analysis分类
农业科技引用本文复制引用
吴凤霞,尤广林,邓广明,孙建平..楸木板材干燥含水率时间序列仿真分析与预测[J].西北林学院学报,2024,39(1):223-227,255,6.基金项目
国家自然科学基金(32260359) （32260359）
广西自然科学基金重点项目(2022GXNSFDA035065). （2022GXNSFDA035065）
