摘要
粉质黏土区域使用钻孔灌注桩进行桥梁桩基础施工需要保证护壁泥浆的失水率在合理范围。为提高泥浆护壁效果,依托现场泥浆生产及室内检测数据,采用麻雀搜索算法(SSA)优化的极端梯度提升算法(XGBoost),建立了SSA-XGBoost钻孔灌注桩泥浆失水率预测模型,并对比了反向传播神经网络(BPNN)、支持向量机(SVM)、XGBoost、SSA-BP和SSA-SVM这5种算法模型的预测性能,以探究不同模型之间的预测性能差异。研究结果表明,SSA-XGBoost模型预测精度最高,均方根误差相较于SSA-BP、SSA-SVM和XGBoost分别减少了58.89%、70.29%和7.58%,预测精度相较于其他模型提升7%以上,模型的精度能够满足现场泥浆质量控制要求,可用于桥梁钻孔灌注桩泥浆配合比的调整。
For the construction of bridge pile foundations using drilled pile in loess areas,it is necessary to ensure that the water loss rate of the retaining mud is controlled within a reasonable range.To improve the effectiveness of mud wall protection,based on on-site mud production and indoor testing data,the extreme gradient boosting algorithm(XGBoost)optimized by sparrow search algorithm(SSA)was used to establish the SSA-XGBoost prediction model for mud water loss rate of drilled pile.The performance prediction of five algorithm models including back propagation neural network(BPNN),support vector machine(SVM),XGBoost,sparrow search algorithm-back propagation network(SSA-BP),and sparrow search algorithm-support vector machine(SSA-SVM)was compared to explore the differences in performance prediction between different models.The research shows that the SSA-XGBoost model has the highest prediction accuracy,root mean square error has decreased by 58.89%,70.29%,and 7.58%compared with SSA-BP,SSA-SVM,and XGBoost,with a prediction accuracy improvement of more than 7%compared with other models.The accuracy of the model can meet the requirements of on-site mud quality control and can be used for adjustment the mud mix ratio of bridge drilled pile.
作者
张文博
ZHANG Wenbo(China Communications Construction Infrastructure Maintenance Group Co.,Ltd.,Beijing 100011,China)
出处
《技术与市场》
2026年第2期77-82,共6页
Technology and Market
