基于优化支持向量机-混沌BP神经网络的基坑变形预测研究

Study of Deformation Prediction of Foundation Pit Based on Optimized Support Vector Machine and Chaotic BP Neural Network

为解决基坑变形预测精度低的问题,采用小波去噪分离基坑变形的趋势项及误差项序列,并利用多种优化的支持向量机对趋势项序列进行预测,采用混沌BP神经网络对误差项序列进行预测,将两者预测结果进行叠加即得到变形预测值,且可根据后期监测数据的更新,实时增加数据信息,达到跟踪预测的目的。经过3个实例检验,得出小波函数的去噪效果相对较优,且预测结果的相对误差均值均小于2%,验证了优化支持向量机-混沌BP神经网络模型的有效性,且该模型具有预测精度高、适用性强等优点,对掌握基坑变形的发展趋势及评价基坑的稳定性具有重要意义。

The accuracy of deformation prediction of foundation pit is low by using traditional methods. As a result, the tendency item and error item sequence of foundation pit deformation are separated by wavelet; the tendency item sequence is predicted by some optimized support vector machines; the error item sequence is predicted by chaotic BP neural network. The deformation prediction results of foundation pit can be obtained by superposition of the two prediction results; and the tracing prediction can be realized by adding later monitoring data uploading. According to case study results, the denoising effect of the wavelet functions are relatively superior and the mean relative error of the prediction results are less than 2%, which verify the validity, prediction accuracy and high adaptability of the optimized support vector machine-chaotic BP neural network model.