基于关联分析和遗传算法优化BP的隧道围岩变形预测

隧道工程处于岩土介质中,岩体自然因素与隧道围岩变形难以用确定的关系表述。因此,通过现场监测隧道变形情况,预测隧道围岩变形具有重要意义。选取我国地势第二阶梯的川陕鄂黔中、低山区,以吴家沟隧道为依托,基于灰色关联分析,选取影响隧道围岩变形的主要因素,基于生物进化的思想,用遗传算法优化BP神经网络,并验证该算法的正确性和精确性。应用工程实际,得到空间维预测结果,为实际应用提供借鉴。结果表明,在隧道围岩变形预测中,遗传算法优化神经网络比原始算法精度高,满足隧道围岩变形预测精度的需要,对长大高风险隧道围岩变形预测有一定的参考意义。

Deformation Prediction of Tunnel Surrounding Rock Based on Correlation Analysis and Genetic-Algorithm Optimized BP

The tunnel engineering is in the geotechnical medium, and the natural factors of the rock mass and the deformation of the surrounding rock of the tunnel are difficult to be expressed in a certain relationship. Therefore, it is of great significance to predict the deformation of tunnel surrounding rock by monitoring the deformation of the tunnel on site. The middle and low mountainous areas of Sichuan, Shaanxi, Hubei and Guizhou, which are the second step of China’s topography, are selected based on the Wujiagou Tunnel. Firstly, based on the grey relational analysis, the main factors affecting the deformation of the surrounding rock of the tunnel are selected. Then, based on the idea of biological evolution, the genetic algorithm is used to optimize the BP neural network, and the correctness and accuracy of the algorithm are verified. Finally, by applying the engineering reality, spatial dimension prediction results are obtained, which provides some reference for practical applications. The results show that in the prediction of tunnel surrounding rock deformation, the genetic algorithm optimization neural network is more accurate than the original algorithm, and it meets the needs of tunnel surrounding rock deformation prediction accuracy, and has certain reference significance for the prediction of surrounding rock deformation of long and high risk tunnels.