基于模糊聚类循环迭代模型的煤仓沉降预测应用研究

Application Research of Coal Bunker Subsidence Prediction Based on Fuzzy Clustering Loop Iteration Model

针对神经网络等模型在时间、荷载等综合因素影响下预测精度不高的问题,提出将模糊聚类循环迭代模型应用于沉降预报,根据前期经验数据对后期沉降趋势进行模拟,引用平均相对误差、均方根误差分别衡量总体精度和偏差;经验证模型精度优于BP神经网络和支持向量机等3种方法。结果表明:基于模糊聚类循环迭代模型适用于多因素影响下的煤仓沉降预测,新沉降预测模型应用也将为工程设计应用提供更多参考。

Aiming at the prediction accuracy of neural network and other model is not high under the influence of time ,load and other comprehensive factors,fuzzy clustering loop iterative model was proposed for coal bunker subsidence prediction. The later subsidence tendency simulated on the basis of preliminary date and average relative error and the root mean square error were used to measure overall accuracy and deviation respectively. It is proved that the model accuracy is better than the three models such as BP neural network and support vector machines. The results show that the fuzzy clustering loop iterative model is suitable for coal bank subsidence prediction under the influence of multiple factors. Also,the application of the new subsidence prediction model will provide more reference for engineering design applications.