不等时距GM(1,1)模型在建筑沉降预测中可行性研究

VERIFICATION ON DIFFERENT INTERVAL GM(1,1)MODEL FOR BUILDING SETTLEMENT PREDICTION

在建筑变形预测中,不等时距 GM(1,1)模型是被广泛引用的经典模型。在数十例工程实践中运用该模型,通过预测与实测成果对比,分析模型的可行性。结果表明,该预测模型存在很大的误差,不能达到预报建筑物最终沉降的要求,建议结合场地条件,建筑结构特征来预测建筑物沉降。

The different interval GM （ 1,1 ） model for building settlement prediction is testified in dozens of practical projects in Zhengzhou. The predicted settlements and field survey ones are compared and analyzed. The result shows that there is a great error in the different interval GM （1,1） model that is established according to the practical measured data, and the predicted model cannot predict the settlement of the building correctly. Therefore, the predicted model cannot be used for building settlement forecasting. Predicted model taking into account the field and structure features should be set up to predict the settlement of building.