时间序列分析方法预报基坑支撑轴力

介绍了自回归模型AR(n)的建立步骤。基坑轴力监测数据波动很大，但是模型预测结果平均误差只有3%左右，说明模型有很强的自适应性，同时证实时间序列分析法可应用于基坑支撑轴力监测数据分析和预报。应用该模型对现场观测数据进行分析，对模型参数进行估计及检验，并判断模型的正确性，在此基础上预报其变化趋势。结果证明，误差较大地方是由于施工的干扰造成的分析，从而可以判断出施工过程对支撑轴力的影响程度。

Time series analysis method for predicting supporting forces of excavation pit

Time series analysis method is applied to analyze the monitoring data of supporting forces of excavation pit.The original auto-regressive model is utilized to search the most suitable model by continuously calculating the relevant parameters.The suitability of the model is determined by checking the residual mean square deviation.If the difference of square deviation between successive models is small enough,the last one should be identified as the most suitable model.On the basis of the most suitable model and corresponding parameter the formula for predicting the supporting forces is deduced.The application of this method shows that the average deviation of prediction is less than 3%,no matter how the monitoring data variation is.It demonstrates that the model possesses good auto-adaptability.