时间序列—马尔科夫组合模型在建筑物沉降变形监测中的应用

Application of Time series-Markov composite Model in the deformation monitoring of building settlement

为了研究时间序列—马尔科夫组合模型在建筑物沉降变形监测中的应用,文中对无偏灰色模型、时间序列线性移动平均法及马尔科夫模型进行研究,并对3种方法的预测结果和精度进行对比分析。结果表明:由于随机性波动的影响,传统的无偏灰色模型预测不易显示出沉降趋势,且预测周期较短。时间序列线性移动平均法和马尔科夫模型可以处理时间序列的随机波动,能克服无偏灰色模型预测随机波动性大的序列时精度较低的问题。结合二者构建的时间序列—马尔科夫组合模型,预测精度高、中长期预测能力强,更适用于建筑物沉降非线性变化的特点,可以为建筑物沉降的中长期预测提供理论支持。

In order to study the application of Time series-Markov composite model in building subsidence deformation monitoring,this paper examines the unbiased grey model,time series linear moving average method and Markov model,in which the prediction outcomes and accuracy of the three methods are compared and analyzed.The conclusion shows that due to the influence of random fluctuations,it is not easy for the traditional unbiased grey model to show the subsidence trend with a short prediction period.In fact,the time linear moving average method and Markov model can not only deal with the random fluctuations of time series,but also overcome the problem of low accuracy of unbiased grey model in the prediction of series with great random fluctuations.It can be concluded that Time series-Markov composite model which combines the two methods is more accurate with higher level of medium and long-term prediction,which is more suitable for the nonlinear change of building subsidence.Moreover,it can provide support for the medium and long-term prediction of building subsidence.