基于POI-ConvLSTM模型的周期来压预测研究  

Prediction of periodic weighting based on POI ConvLSTM model

作　　者：尹春雷 YIN Chunlei(Beijing Tianma Intelligent Control Technology Co.,Ltd.,Beijing 101399,China;School of Software,Beihang University,Beijing 100191,China)

机构地区：[1]北京天玛智控科技股份有限公司,北京101399 [2]北京航空航天大学软件学院,北京100191

出　　处：《煤炭工程》2024年第9期121-126,共6页Coal Engineering

摘　　要：针对综采工作面周期来压预测的技术难题,研究了理论分析、数据采集与预处理、模型评估与优化等方法,提出了具有时空关联分析与POI(Point of Intersesting)数据的ConvLSTM模型,利用多源数据融合得到周期来压预测的最优解,实现工作面环境状态的实时感知和预测。试验结果表明:基于POI-ConvLSTM的工作面周期来压预测模型,均方误差为0.159,R 2评价指标为0.999,相比于Seq2Seq和ConvLSTM模型的均方误差分别降低了68.07%和4.22%。可见,融合了多元数据POI-ConvLSTM模型的预测精度更高,普适性更强,能够准确地提前预测周期来压问题。Aiming at solving the technical difficulties in predicting the periodic weighting of fully mechanized mining face,theoretical analysis,data acquisition and pre-processing,model evaluation and optimization are studied,and the ConvLSTM model with spatio-temporal correlation analysis and POI(Point of Intersesting)data is proposed.Based on multi-source data consolidation,the optimal solution for periodic weighting prediction is obtained and real-time sensing and prediction of working face environment is realized.The test results show that the mean square error based on POI ConvLSTM model for working face periodic weighting is 0.159,and R 2 evaluation index is 0.999.Compared with the Seq2Seq and ConvLSTM models,the mean square error is reduced by 68.07%and 4.22%,respectively.Therefore,the POI ConvLSTM model with multiple data has more precise prediction and universality,which will accurately predict periodic pressure in advance.

关 键 词：POI ConvLSTM 周期来压 时空关联 

分 类 号：TD323[矿业工程—矿井建设]