金华市某岩质滑坡成因机制分析及发展趋势预测

为探究金华市某岩质滑坡成因机制,多方面分析表明该滑坡是在固有内部因素条件下由地下水活动诱发产生的。为评价滑坡发展趋势,解决陈旧信息和波动性数据造成的传统灰色GM(1,1)模型预测精度较低的问题,提出一种利用动态新陈代谢和马尔科夫模型对原始灰色GM(1,1)模型改进的方法。通过新陈代谢理论来实现数据的动态更新,使得灰色GM(1,1)动态模型的预测值更加接近最新的变化趋势。利用马尔科夫模型对得到的灰色动态新陈代谢GM(1,1)模型预测值进行修正,提高了模型的预测精度。预测结果表明,灰色-马尔科夫动态新陈代谢GM(1,1)模型在滑坡形变预测中的预测平均相对误差相比于传统的灰色GM(1,1)模型降低了70%。对于波动性较大的滑坡监测数据,灰色-马尔科夫动态新陈代谢GM(1,1)模型预测精度远优于传统灰色GM(1,1)模型,具有实际的参考价值。金华市某岩质滑坡监测数据和预测结果研究表明,该滑坡发展趋势不稳定,需采取适当的防治措施。

Genetic mechanism analysis and development trend prediction of a rock landslide in Jinhua

A multi-analysis has been conducted to explore the genetic mechanism of a rock landslide in Jinhua City, indicating that the landslide was induced by groundwater movement with the inherent internal factors. An improved Grey Model GM(1,1) using Dynamic Metabolism and Markov Model is proposed to evaluate the landslide development trend and increase the low prediction accuracy of original GM(1,1) caused by staled information and volatility data. The predicted value of Dynamic GM(1,1) is closer to the latest variation trend owing to the dynamic update of data realized by Metabolic Theory. The prediction accuracy is also improved by utilizing Markov Model to correct the predicted value of Dynamic Metabolism GM(1,1). The prediction result shows that the average relative error of prediction of Markov Dynamic Metabolism GM(1,1) is reduced by 70% in the deformation prediction of landslides, compared with the traditional Grey Model (1,1). The Markov Dynamic Metabolism GM (1,1) possesses a far superior prediction accuracy to the traditional GM (1,1) in the matter of the landslide monitoring data with large fluctuation, thus it has practical reference value. Considering that the development trend of the landslides is unstable according to the monitoring data and prediction results in Jinhua City, appropriate control measures should be taken.