基于概率积分法和自四归模型的开采沉陷算法

The Subsidence Prediction of Goaf Based on Probability Integral Method and Auto Regressive Moving Average Model

工程地质条件复杂的采空区,采用概率积分法预测地表沉陷量时会出现较大偏差,因而提出概率积分法和ARMA(auto-regressive and moving average model)模型相结合的采空区地表沉陷预测模型。模型将采空区地表沉陷分为两部分:受采空区影响的趋势项沉陷和受工程地质条件影响的突变项沉陷;利用平动法分离这两项沉陷量,然后分别采用概率积分法和ARMA模型对趋势项和突变项进行沉陷分析预测,两项之和为采空区最终沉陷预测结果。采用该模型对湖南省冷水江宝大兴矿区采空区地表沉陷预测,预测结果较好地反映出复杂工程地质条件作用下采空区地表沉陷的变化发展趋势。通过对监测值、概率积分法预测值和新模型预测值进行对比,新模型预测精度较高,与监测值相吻合,在相似的复杂条件下具有较高的可行性。

In view of the complicated engineering geological conditions of goaf area,it occurs large deviation when the conventional probability integral method is adopted to predict the subsidence quantity,thus a new method of mining subsidence prediction with the combination of the model based on probability integral method and the auto regressive model is put forward.The mining subsidence can be divided into two parts：trend term subsidence affected by mining and mutation subsidence affected by the engineering geological conditions in the new method,and they may be separated with the use of translation method,then the two parts can be figured out respectively by the probability integral method and the index model,and the sum of them is the final subsidence prediction result.Taking the new method at Lengshui River Bao daxing mining goaf in hunan province for example,the calculation results are able to meglio embody the trend of the development of the displacement under the action of the complicated engineering geological conditions in mining area.The predicted effect and accuracy of the new method is preferable,and the results are in conformity with the actual results,hence it has a high feasibility in similar complicated conditions.