中文摘要：

在收集肥城煤田封堵突水点资料基础上,分析了影响注浆量的主要因素是突水水压、突水量、封堵过水通道长度、注浆压力等,借助智能算法自动获取支持向量机最佳参数的优点,优化支持向量机回归分析能力,建立GA-SVR非线性模型和PSO-SVR非线性模型,并通过实际工程对封堵突水点注浆量做出预测。通过对比实际注浆预测结果,得出PSO-SVR模型预测结果相对准确,但预测结果波动性偏大,GA-SVR预测结果相对稳定,但预测结果误差相对偏大的特点。因此提出在进行注浆量预测时,采取两种模型同时进行注浆量预测,取其区间值,实现又快又好又经济地封堵突水点。

英文摘要：

Based on the data of blocking water inflow point of Feicheng coalfield, this paper analyzed the main factors which affect grouting quantity, including water pressure, water inrush, length of sealing off water channel, grouting pressure, and the optimal param eters of the support vector machine was figured out by the intelligent algorithm. The GA-SVR nonlinear model and PSO-SVR model were established. By comparing the results of actual grouting, the PSO SVR model was more accurate than the GA-SVR model, and the GA SVR forecasting results were more stable and less volatile. Therefore, it would better use two kinds of models to predict the amount of grouting at the same time and take the interval value, so as to achieve fast and favorable sealing at water inrush point.