基于宽度学习的结构沉降预测

结构监测是确保工程结构建设在施工和运营阶段安全的关键因素,因此采用合理有效的预测模型对结构沉降监测数据进行科学准确的预测成为了当前结构沉降预测研究的重点。针对传统预测方法与深度学习方法用于结构沉降预测存在的预测精度不够高、模型结构复杂、训练耗时等问题,提出了一种基于宽度学习的结构沉降时间序列预测模型。通过实测地铁地下隧道沉降监测数据对宽度学习、人工神经网络、支持向量回归和深度置信网络-支持向量回归预测模型的预测结果进行对比分析。实验结果表明:宽度学习系统(broad learning system,BLS)应用于结构沉降预测具有良好的效果,其训练速度更快,预测精度更高。验证了所提出的宽度学习算法应用于结构沉降预测的可实施性和有效性。

Structural Settlement Prediction Based on Broad Learning

Structural monitoring is a key factor to ensure the safety of the engineering structure during the construction and operation stages.The scientific and accurate prediction of structural settlement monitoring data by using reasonable and effective prediction models has become the focus of current structural settlement prediction research.In view of the problems existing in the traditional prediction methods and deep learning methods for structural settlement prediction,such as low accuracy,complex model structure and training time consuming,etc,we propose a structure settlement time series prediction model based on broad learning.The prediction results of the broad learning,artificial neural network,support vector regression and deep confidence network-support vector regression prediction model are compared and analyzed by the measured settlement monitoring data of subway tunnel.The experiment shows that the broad learning system(BLS)is effective on the prediction of structural settlement with faster training speed and higher prediction accuracy.The feasibility and effectiveness of the proposed broad learning algorithm proposed for structural settlement prediction are verified.