组合预测模型在建筑物沉降预测中的应用

Application of Combined Prediction Model in Building Settlement Prediction

分析预测建筑物的未来沉降变化规律及发展趋势能够有效保障建筑物在施工及使用期间的安全。现有的时间序列模型、灰色GM(1,1)模型、线性拟合模型等单一预测模型均存在不足之处,预测精度不理想。通过挖掘各单一预测模型的有效信息,组合各单一预测模型优势,采用经典定权法建立组合预测模型来预测建筑物未来的沉降变化和发展趋势。以芜湖中建生活广场4#楼第5~10期的沉降监测数据为基础,预测了4#楼第11期的沉降变化规律及其发展趋势。预测结果显示,组合预测模型的预测精度更高,效果更好,具有一定的借鉴意义与应用前景。

The analysis and prediction of the future settlement change law and development trend of buildings can effectively guarantee the safety during building construction and use. There are shortcomings in the existing time series model, grey GM(1,1) model, linear fitting model and other single prediction models, and so the ideal prediction accuracy cannot be achieved. By digging the effective information of each single prediction model and combining the advantages of each single prediction model, a combined prediction model is established by using the classical fixed weight method to predict the future settlement change and development trend of buildings. Based on the settlement monitoring data of phase 5~10 of No.4 building in Wuhu China Construction living square, the settlement change law and development trend of phase 11 of building No.4 are predicted. The prediction results show that the combination prediction model has higher prediction accuracy and better effect, which has certain reference significance and application prospect.