基于FOA-SVR的矿井底板突水量预测模型应用研究

矿井突水是常见的突发性强烈的矿井灾害。为了更好地预防矿井水灾,降低灾害造成的物质损失以及减少人员伤亡,建立了一种基于FOA-SVR的矿井底板突水量预测模型,利用果蝇算法优化支持向量回归机算法(FOA-SVR)选出最优的模型参数。针对底板突水这种非线性、小样本问题,从突水因素中选取水压、含水层厚度、隔水层厚度、底板采动裂隙带深度以及断层落差这5个作为特征因素。然后利用FOA对SVR参数进行优化之后建立FOA-SVR底板突水量预测模型,输出即为需要预测的突水量。结合实例并将该模型的预测结果与SVR模型的预测结果进行对比,结果表明：该模型在预测突水量的精度上比SVR模型更高,具有一定的应用价值。

Application of the prediction model for mine floor water inrush quantity based on FOA-SVR

Mine water inrush is a common mine disaster which is sudden and intense. In order to better prevent mine floods and reduce material losses and casualties caused by disasters, a prediction model for mine floor water inrush quantity was built by studying the principle and characteristics of FOA-SVR, using the fruit fly algorithm (FOA) to optimize support vector regression algorithm (SVR) to select the excellent model parameters. For the nonlinear, small sample problem of floor water inrush, water pressure, aquifer thickness, aquiclude thickness, height of water flowing fractured and fault throw were selected as the characteristic factors. After the SVR parameters were optimized by the fruit fly algorithm, the FOA-SVR prediction model for floor water inrush quantity was built, and the output was the water inrush quantity that we need to predict. Comparing the predicted results of FOA-SVR model with the predicted results of SVR model, the results shows that, the model is more accurate than the SVR model in the accuracy of predicting water inrush quantity and has certain application value.