基于人工神经网络模型的木材干燥应变模拟预测

doi:10.11707/j.1001-7488.20200608

摘要/Abstract

摘要：

目的: 研究常规干燥过程中干燥基准、预处理条件、含水率对木材干燥应力的影响，探讨干燥应力沿髓心至树皮方向的分布情况，以实现干燥应变的模拟预测。方法: 整合分析采用图像解析法测算得到的弹性应变和机械吸附蠕变相关数据，基于人工神经网络模型，以干燥温度、含水率、相对湿度、距髓心距离为输入变量对弹性应变进行模拟预测，以预处理温度、干燥温度、含水率、相对湿度、距髓心距离为输入变量对机械吸附蠕变进行模拟预测。通过网络训练和验证，得到合理的人工神经网络预测模型，并对模型进行测试，探讨分析所建立模型的预测能力。结果: 弹性应变预测模型中，各数据集均呈现出较好的相关性，训练集、验证集和测试集的相关系数（R）分别为0.988、0.983和0.978，所有数据集的决定系数（R²）均高于0.95，验证集达到最优时的均方差（MSE）为1.21×10^-6。机械吸附蠕变预测模型中，利用含水率为28%和12%的数据集进行模型训练和验证，训练集和验证集的相关系数（R）分别为0.981、0.977，验证集达到最优时的均方差（MSE）为1.26×10^-6；利用含水率20%的数据集进行模型测试，测试集的相关系数（R）为0.969，所有数据集的决定系数（R²）均高于0.94，网络模型能够解释94%以上的试验数据，表现出较好的预测能力。结论: 所建立模型的预测值和试验值吻合较好，预测成功率较高，能够为人工神经网络在干燥应力、应变方面的应用提供可行性依据。

关键词: 人工神经网络, 弹性应变, 机械吸附蠕变, 模拟预测

Abstract:

Objective: The effects of drying schedule, pretreatment condition, moisture content on wood drying stress and its distribution from pith to bark were investigated in this study to achieve the simulation and prediction of drying strain. Methods: The data sets of elastic strain and mechano-sorptive creep obtained with image analysis method were analyzed. The elastic strain and mechano-sorptive creep were modelled by artificial neural network, for the model of elastic strain with four inputs, i.e., drying temperatures, wood moisture content, relative humidity and the distance from pith, and for mechano-sorptive creep added pre-steaming temperature as an additional input. According to the training and validation processes, two reasonable prediction models were obtained, and then the predictive ability of the models were discussed with testing processes. Results: In elastic strain model, there were significant correlations between experimental and predicted values in all date sets. The R-values for training, validation and test sets were 0.988, 0.983 and 0.978, respectively. The R² values were greater than 0.95 in all data sets, and the best validation performance of the mean square error was 1.21×10^-6. In mechano-sorptive creep model, the R-values for training and validation sets were 0.981 and 0.977 with the data sets at moisture content of 28% and 12%, respectively. The best validation performance of the mean square error was 1.26×10^-6. The R-value for test set was 0.969 with the data set at moisture content of 20%. Furthermore, the R² values were greater than 0.94 in all data sets, indicating that the network model was capable to explain more than 94% experimental values. Conclusion: In the two established models, the predicted values were in good agreement with the experimental values, showing a high prediction accuracy, which provided a feasible basis for the application of artificial neural network in exploring drying stress and strain.

Key words: artificial neural network, elastic strain, mechano-sorptive creep, simulation and prediction

中图分类号:

S781.71

付宗营,蔡英春,高鑫,周凡,江京辉,周永东. 基于人工神经网络模型的木材干燥应变模拟预测[J]. 林业科学, 2020, 56(6): 76-82.

Zongying Fu,Yingchun Cai,Xin Gao,Fan Zhou,Jinghui Jiang,Yongdong Zhou. Simulation of Drying Strain Based on Artificial Neural Network Model[J]. Scientia Silvae Sinicae, 2020, 56(6): 76-82.

图/表 11

图1

图2

表1

图3

图4

表2

图5

图6

图7

图8

图9

参考文献 0

	付宗营, 赵景尧, 蔡英春. 树盘干缩异向性引起应变的测算及分析. 中国工程科学, 2014. 164, 25- 29. doi: 10.3969/j.issn.1009-1742.2014.04.005
	Fu Z Y , Zhao J Y , Cai Y C . Investigation of tangential strain caused by shrinkage anisotropy using image analytical method. Engineering Sciences, 2014. 164, 25- 29. doi: 10.3969/j.issn.1009-1742.2014.04.005
	李坚. 木材科学. 北京: 科学出版社. 2014. 254
	Li J . Wood science. Beijing: Science Press. 2014. 254
	蒋佳荔, 吕建雄. 木材干燥应力的研究方法与进展. 木材工业, 2005. 192, 4- 7.
	Jiang J L , Lü J X . Review of research methods for wood drying stresses. China Wood Industry, 2005. 192, 4- 7.
	杨文斌, 陈眉雯. 利用神经网络预测木材径向导热系数. 林业科学, 2006. 423, 25- 28.
	Yang W B , Chen M W . Predicting the wood radial thermal conductivity using neural network. Scientia Silvae Sinicae, 2006. 423, 25- 28.
	Avramidis S , Iliadis L . Predicting wood thermal conductivity using artificial neural networks. Wood and Fiber Science, 2005a. 37, 682- 690.
	Avramidis S , Iliadis L . Wood-water sorption isotherm prediction with artificial neural networks:a preliminary study. Holzforschung, 2005b. 59, 336- 341. doi: 10.1515/HF.2005.055
	Avramidis S , Iliadis L , Mansfield S D . Wood dielectric loss factor prediction with artificial neural networks. Wood Science and Technology, 2006. 40, 563- 574. doi: 10.1007/s00226-006-0096-3
	Avramidis S , Wu H W . Artificial neural network and mathematical modeling comparative analysis of nonisothermal diffusion of moisture in wood. HolzalsRoh-und Werkstoff, 2007. 65, 89- 93. doi: 10.1007/s00107-006-0113-0
	Bedelean B , Lazarescu C , Avramidis S . Predicting RF heating rate during pasteurization of green softwoods using artificial neural networks and Monte Carlo method. Wood Research, 2015. 601, 83- 94.
	Ceylan I . Determination of drying characteristics of timber by using artificial neural networks and mathematical models. Drying Technology, 2008. 26, 1469- 1476. doi: 10.1080/07373930802412132
	Esteban L G , Fernandez F G , Palacios P D . MOE prediction in Abies pinsapo Boiss. timber:application of an artificial neural network using non-destructive testing. Computer and Structure, 2009. 87, 1360- 1365.
	Fu Z Y , Zhao J Y , Huan S Q , et al. The variation of tangential rheological properties caused by shrinkage anisotropy and moisture content gradient in white birch disks. Holzforschung, 2015. 695, 573- 579.
	Hagan M T , Demuth H B , Beale M H . Neural network design. Boston: PWS Publishing Company. 1996.
	Iliadis L , Mansfield S D , Avramidis S , et al. Predicting Douglas-fir wood density by artificial neural networks ANN based on progeny testing information. Holzforschung, 2013. 677, 771- 777.
	Mansfield S D , Iliadis L , Avramidis S . Neural network prediction of bending strength and stiffness in western hemlock Tsuga heterophylla Raf. Holzforschung, 2007. 616, 707- 716.
	Myhara R M , Sablani S . Unification of fruit water sorption isotherm using artificial neural networks. Drying Technology, 2001. 198, 1543- 1554.
	Pang S . Modelling of stress development during drying and relief during steaming in Pinus radiata lumber. Drying Technology, 2000. 188, 1677- 1696.
	Sarle W S . Stopped training and other remedies for overfitting. Proceedings of the 27^th Symposium on the Interface of Computing Science and Statistics, 1995. 352-360
	Tiryaki S , Hamzacebi C . Predicting modulus of rupture MOR and modulus of elasticity MOE of heat treated woods by artificial neural networks. Measurement, 2014a. 49, 266- 274. doi: 10.1016/j.measurement.2013.12.004
	Tiryaki S , Aydın A . An artificial neural network model for predicting compression strength of heat treated woods and comparison with a multiple linear regression model. Construction and Building Materials, 2014b. 62, 102- 108. doi: 10.1016/j.conbuildmat.2014.03.041
	Watanabe K , Matsushita Y , Kobayashi I , et al. Artificial neural network modeling for predicting final moisture content of individual Sugi Cryptomeria japonica samples during air-drying. Journal of Wood Science, 2013. 59, 112- 118. doi: 10.1007/s10086-012-1314-2
	Watanabe K , Kobayashi I , Matsushita Y , et al. Application of near-infrared spectroscopy for evaluation of drying stress on lumber surface:a comparison of artificial neural networks and partial least squares regression. Drying Technology, 2014. 325, 590- 596.
	Zhan J F , Avramidis S . Mechano-sorptive creep of hemlock under conventional drying:ii. description of actual creep behavior in drying lumber. Drying Technology, 2011. 2910, 1140- 1149.
	Zhang D Y , Sun L , Cao J . Modeling of temperature-humidity for wood drying based on time-delay neural network. Journal of Forestry Research, 2006. 172, 141- 144.
	Zhang D Y , Liu Y X , Cao J , et al. Neural network prediction model of wood moisture content for drying process. Scientia Silvae Sinicae, 2008. 4412, 94- 98.

网络训练误差 Network training error	神经元数量Number of neurons
网络训练误差 Network training error	3	4	5	6	7	8	9
M1	2.83×10^-6	1.97×10^-6	1.56×10^-6	1.21×10^-6	9.67×10^-7	—	—
M2	3.56×10^-6	2.71×10^-6	2.23×10^-6	1.16×10^-6	1.87×10^-6	1.12×10^-6	9.20×10^-7

组别Groups	线性回归方程Linear regression equation	R	R²
训练集Training set	Pre=0.98Exp-7.7e-05	0.988	0.97
验证集Validation set	Pre=0.95Exp-0.000 31	0.983	0.96
测试集Test set	Pre=1Exp+0.000 1	0.978	0.95
总数据集All data set	Pre=0.98Exp-9.1e-05	0.985	0.97

[1]	王强;胡海清. 基于岭回归和人工神经网络估测森林可燃物负荷量[J]. 林业科学, 2012, 48(9): 108-114.
[2]	杨广斌;刘鹏举;唐小明. 动态数据驱动的林火蔓延模型适宜性选择[J]. 林业科学, 2011, 47(1): 107-112.
[3]	虞华强赵荣军刘杏娥费本华. 木材蠕变模拟研究概述[J]. 林业科学, 2007, 43(7): 101-105.
[4]	胡林冯仲科聂玉藻. 基于VLBP神经网络的林火预测研究[J]. , 2006, 42(zk): 155-158.
[5]	齐巍王立海. 基于小波神经网络的木材内部缺陷类型识别的研究[J]. 林业科学, 2006, 42(8): 63-68.
[6]	杨景标马晓茜. 基于人工神经网络预测广东省森林火灾的发生[J]. 林业科学, 2005, 41(4): 127-132.
[7]	杨文斌刘一星刘迎涛. 人工神经网络在木材工业中的应用前景[J]. 林业科学, 2004, 40(6): 153-157.
[8]	战剑锋顾继友艾沐野. 白桦木材干燥过程横纹流变特性的初步研究[J]. 林业科学, 2004, 40(5): 174-179.
[9]	赵茂程郑加强林小静封晓强. 基于BP神经网络的树形识别系统研究[J]. 林业科学, 2004, 40(1): 154-157.
[10]	何东进洪伟吴承祯. 毛竹林林分平均胸径模拟预测模型的研究[J]. 林业科学, 2000, 36(zk): 148-153.
[11]	李大纲顾炼百. 杨木高温干燥过程中表层流变特性的研究[J]. , 1999, 35(1): 83-89.