支持向量机模型在电站闸坝沉降变形预测中的应用

Application of support vector machine model in prediction of the settlement deformation of the hydropower station dam

本文结合支持向量机模型,建立多元自回归方程对辽宁东部某电站闸坝沉降变形进行预测,并结合闸坝沉降变形监测点监测的沉降数据对模型进行验证。结果表明:支持向量机模型可用于电站闸坝沉降变形预测,模型预测的沉降值和监测点沉降监测值年尺度和月尺度的相对误差均在20%以内,预测沉降值和监测沉降值之间的绝对误差小于1.5mm,且支持向量机模型可准确预测闸坝沉降的变化趋势。研究成果对于闸坝除险加固中沉降预测提供较好的方法参考。

In this paper,a multivariate autoregressive equation is established to predict the settlement and de- formation of a dam in Liaoning Province,and the model is verified by the subsidence data monitored at the site of deformation monitoring.The results show that the support vector machine model can predict the settlement and deformation of the dam well.The relative error of the settlement value and the settlement value is less than 20% on the annual and monthly scale.The predicted settlement value and the monitoring settlement value and the absolute error is less than 1.5mm,the support vector machine model can accurately predict the trend of dam settlement.The research results provide a good reference for the settlement prediction of the dam rein- forcement.