基于松散型小波神经网络的某隧道仰坡变形趋势研究

针对某隧道洞口仰坡的稳定性问题,首先选取松散型小波神经网络对仰坡的沉降变形进行预测,再利用秩相关系数检验和Mann-Kendall检验对仰坡的变形趋势进行判断;结果表明:不同小波函数的去噪效果不一致,其中监测点#5采用db4小波的去噪效果最好,而监测点#6则是采用db7小波的去噪效果最好;另外,秩相关系数检验和Mann-Kendall检验在仰坡变形趋势的判断中具有很好的一致性,均得出两监测点的变形趋势是往减小方向发展,仰坡变形趋于稳定,可以安全地进行后期施工。

Research on Deformation Trend of Tunnel Upward Slope Based on Loose-type Wavelet Neural Network

Aiming at the stability of upward slope at the tunnel entrance, the loose-type wavelet neural network is selected to predict the settlement deformation; the rank correlation coefficient test and Mann-Kendall test are performed to judge the deformation trend of upward slope. The results show that the denoising effect of different wavelet functions is inconsistent; wherein, the denoising effect of db4 wavelet at monitoring points #5 is the best, and the denoising effect of db7 wavelet at the monitoring point #6 is the best. In addition, the rank correlation coefficient test and the Mann-Kendall test have very good consistency in the judgment of the deformation trend of upward slope. The deformation trend of two monitoring points is gradually decreased; and the deformation of upward slope is stable, so later construction can be done in a safe manner.