摘要
为研究砂岩在水和温度作用下的抗压强度特性,以三峡库区砂岩为研究对象,进行温度场、渗流场、应力场耦合试验研究,建立了粒子群优化BP神经网络(PSO-BPNN)预测模型,该模型考虑了影响砂岩抗压强度的多种因素(温度、孔隙水压等),预测砂岩三轴抗压强度值.较传统BP神经网络(BPNN)模型,PSO-BP神经网络模型能够更好地预测三场耦合作用下砂岩的抗压强度变化特征,预测精度更高.
The sandstone in the Three Gorges Reservoir area is used to conduct a thermal-hydraulicmechanical(THM)coupling test to study the compressive strength properties of sandstone under water and temperature effects.Then,aparticle swarm optimization BP neural network(PSO-BPNN)prediction model that considers various influencing factors,such as temperature and pore water pressure,is established to predict the triaxial compressive strength of sandstone.Compared with the traditional BP neural network(BPNN)model,the PSO-BP neural network model can better predict the characteristics of the compressive strength change of sandstone under the THM coupling;and the prediction accuracy is higher.
引文
[1]闻磊,李夕兵,唐海燕,等.变温度区间冻融作用下岩石物理力学性质研究及工程应用[J].工程力学,2017,34(5):247-256.
[2]戎虎仁,白海波,王占盛.不同温度后红砂岩力学性质及微观结构变化规律试验研究[J].岩土力学,2015,36(2):463-469.
[3]彭守建,谭虎,许江,等.不同孔隙水压条件下完整砂岩剪切力学特性试验研究[J].岩石力学与工程学报,2017,36(S1):3131-3139.
[4]刘立峰,孙赞东,韩剑发,等.量子粒子群模糊神经网络碳酸盐岩流体识别方法研究[J].地球物理学报,2014,57(3):991-1000.
[5]邓传军,欧阳斌,陈艳红.一种基于PSO-BP神经网络的建筑物沉降预测模型[J].测绘科学,2018(6):1-7.
[6]胡军,董建华,王凯凯,等.边坡稳定性的CPSO-BP模型研究[J].岩土力学,2016,37(S1):577-582,590.
[7]Avunduk E,Tumac D,Atalay A.Prediction of Roadheader Performance by Artificial Neural Network[J].Tunnelling and Underground Space Technology Incorporating Trenchless Technology Research,2014,44.