论文部分内容阅读
基于不同干湿循环作用下砂岩单轴压缩试验结果,分析了干湿循环效应对砂岩变形、强度、破坏特征等力学特性的影响规律,认为随着干湿循环次数的增加,弹性模量及峰值强度均有降低的趋势,降低幅度先大后小,而且会以某一具体值为临界值发展变化,就本次试验而言为干湿循环20次时的抗压强度和弹性模量值;砂岩的破坏特征亦会受到干湿循环试验次数的影响,干湿次数越少,脆性破坏越明显,反之则延性特征增强,即砂岩会呈现一种从脆性到延性转化的破坏规律。以应变和干湿试验次数为输入层,应力为输出层,构建了2-12-1的三层神经网络本构模型。通过样本的学习与检验,证明该模型能较好地描述干湿循环作用下砂岩的力学性能,验证了利用神经网络方法建立岩石本构模型的可行性与可靠性。基于预测结果,建立了完整的考虑干湿循环效应的砂岩力学特性变化规律的数学函数关系式。
Based on the uniaxial compression tests of sandstone under different wetting and drying cycles, the influence of wetting and drying cycles on the mechanical properties such as deformation, strength and failure characteristics of sandstone is analyzed. It is concluded that with the increase of the number of wetting and drying cycles, the elastic modulus and peak The strength decreases with the decrease first and then small, and will change with a specific value as the critical value. For this test, it is the compressive strength and elastic modulus at 20 times of the wet-dry cycle. The failure characteristics of sandstone are also affected by the number of wetting and drying cycles. The smaller the number of wet and dry cycles, the more obvious the brittle failure is, whereas the ductile characteristic is enhanced. That is to say, sandstone will exhibit a failure rule from brittleness to ductile transformation. Taking the times of strain and wet-dry test as input layer and stress as output layer, a three-layer neural network constitutive model of 2-12-1 was constructed. Through the study and test of samples, it is proved that the model can well describe the mechanical properties of sandstone under wetting and drying cycles. The feasibility and reliability of using neural network to establish the constitutive model of rock are verified. Based on the prediction results, a complete mathematics function formula for the variation of mechanical properties of sandstone considering wetting and drying cycles was established.