基于灰色-马尔柯夫链模型的建筑物变形分析与预报

Deformation analysis and prediction for buildings based on Grey-Markov chain model

变形监测数据的分析和预报一直是建筑物施工及运营管理中的重要内容,其预报的可靠性在很大程度上取决于正确选择科学的预测方法。本文在分析现有建筑变形预测方法的基础上,针对灰色系统理论中的GM(1,1)预测的局限性,通过状态划分及计算马尔柯夫转移概率矩阵的方法,提出了适合于建筑物变形预测的灰色-马尔柯夫链模型,同时对该模型进行了实例验证和分析。研究结果表明,灰色-马尔柯夫链模型综合了灰色预测和马尔柯夫链预测两种模型的优点,大大提高了预测的准确度,为建筑物变形监测的数据处理与预报提供了一种更为科学和有效的方法。

Deformation trend prediction is one of the most important aspects in both the construction of and operation of buildings,and the reliability of deformation prediction depends largely on the choice of prediction method.This paper discussed the existing methods of building deformation prediction.Aiming at the limitation of the traditional grey GM（1,1） prediction model,an improved model is proposed through classifying system states and computing state transition probability matrix.A real engineering case is presented to demonstrate and analyze the new method.The result shows that the improved Grey-Markov chain model is much better than the GM（1,1） in both the fitting accuracy and the prediction efficiency.And the proposed Grey-Markov chain model is very suitable to be applied for the building deformation analysis and prediction.