基于叠加马尔科夫链的边坡位移预测研究

针对边坡监测数据的时效性问题,运用叠加马尔科夫链对边坡位移增量进行预测,选取符合马尔科夫链数据要求的位移增量作为预测依据,采用均值-方差法将数据划分为3个状态分级数据区间,并对2019年11月1日至2019年12月30日的60个数据进行分级,得到了不同步长的转移矩阵和均值向量,预测了2019年12月31日、2020年1月1日和2020年1月2日的位移增量分别为2.88、2.84和2.73,预测精度可达到97.3%、89.08%和89.75%,预测精度高、效果好,为边坡位移监测数据发挥预警作用提供了案例支撑,对处理矿山边坡监测数据具有参考意义。

Research on slope displacement prediction based on superposition Markov chain

Aiming at the issue of timeliness of slope monitoring data, the superimposed Markov chain is used to predict the slope displacement increment, the displacement increment that meets the Markov chain data requirements is selected as the prediction basis, and the mean-variance method is used to divide the data into There are 3 state classification data intervals, and 60 data from November 1, 2019 to December 30, 2019 are classified, and the non-synchronized long transition matrix and mean vector are obtained. It is predicted that December 31, 2019, The displacement increments on January 1, 2020 and January 2, 2020 are 2.88, 2.84 and 2.73, and the prediction accuracy can reach 97.3%, 89.08% and 89.75%. The prediction accuracy is high and the effect is good. It is the slope displacement monitoring data The early warning function provides case support, which is of reference significance for processing mine slope monitoring data.