【摘要】 高斯过程是一种最近发展起来新的机器学习技术,对处理非线性复杂问题具有很好的适应性。岩体爆破效应与其影响因素之间是复杂的非线性关系,针对传统方法的局限性,提出一种基于高斯过程的岩体爆破效应预测的新方法,建立相应的岩体爆破效应预测模型,并应用于三峡工程坝区岩体爆破振动速度、爆破损伤深度与损伤半径的预测。通过三峡现场爆破试验数据,建立训练数据集和测试数据集,采用高斯过程方法建立爆破效应与影响因素之间的各影响因素之间的非线性映射关系。研究结果表明,岩体爆破振动速度、爆破损伤深度与损伤半径的预测结果与现场试验结果比较吻合,用高斯过程方法预测岩体爆破效应是科学可行的。与神经网络方法相比,高斯过程方法具有算法参数自适应化的特点,且适用于小样本问题,预测精度高,并易于实现,具有良好的工程应用前景。更多还原

【Abstract】 Gaussian process(GP) is a newly developed machine learning technology based on statistical theoretical fundamentals.It has become a powerful tool for solving highly nonlinear problems.Conventional methods for forecasting of blasting effect in rock mass often meet great difficulty since relationship between blasting effect and its influencing factors is highly complicated nonlinear one.A new method based on GP is proposed for forecasting of blasting effect in rock mass.The method is applied to blasting engineering of the Three Gorges project in China for forecasting of vibration speed,damage depth and damage radius in rock mass.The field experiments of rock blasting are preformed to obtain the training samples and test samples.Nonlinear mapping relationship between blasting effect and its influencing factors can be constructed by GP learning with the training samples.The prediction results for vibration speed,damage depth and damage radius in rock mass using the method are in good agreement with observations.The results of case studies show that the method is feasible,effective and simple to implement for forecasting of blasting effect prediction in rock mass.It has merits of self-adaptive parameters determination and better capacity for solving nonlinear small sample problems comparing with the artificial neural networks method.The good performance of GP model makes it very attractive for a wide range of application in geotechnical engineering.更多还原