基于Spark Streaming流回归的煤矿瓦斯浓度实时预测

发布时间：2018-08-06 10:36

【摘要】：为了实时分析瓦斯监测流数据并对瓦斯浓度进行准确预测以实现瓦斯灾害实时预警,以实时流数据处理框架Spark Streaming构建基于流回归的瓦斯浓度实时预测系统。系统采用分布式流处理技术,可使基于回归算法的瓦斯浓度预测模型更新周期达到秒级,提高了瓦斯浓度预测精度,满足流式大数据处理的实时性要求。实验表明:应用Spark Streaming流回归预测系统在采样周期为5s的瓦斯监测数据流上进行实时预测时,预测平均均方根误差随模型更新周期的缩短而减小,模型更新周期可达15s,且更新周期为45s时预测总均方根误差最小,既能保证预测精度,又能提高瓦斯灾害预警时效。
[Abstract]:In order to analyze gas monitoring flow data in real time and accurately predict gas concentration in order to realize real time early warning of gas disaster, a real time prediction system of gas concentration based on flow regression was constructed by real time flow data processing framework (Spark Streaming). By using distributed flow processing technology, the renewal period of gas concentration prediction model based on regression algorithm can reach second order, which improves the accuracy of gas concentration prediction and meets the real-time requirements of flow big data processing. The experimental results show that the mean root mean square error (RMS) of the prediction decreases with the shortening of the renewal period of the model when the Spark Streaming flow regression prediction system is used for real-time prediction on the gas monitoring data stream with a sampling period of 5 s. The model updating period can be up to 15 s and the prediction RMS error is the smallest when the renewal period is 45 s, which can not only guarantee the prediction accuracy, but also improve the early warning time of gas disaster.
【作者单位】：中南大学资源与安全工程学院;湖南科技大学计算机科学与工程学院;湖南科技大学资源环境与安全工程学院;湖南科技大学煤炭资源清洁利用与矿山环境保护湖南省重点实验室;
【基金】：煤矿安全开采技术湖南省重点实验室开放基金项目(201304) 湖南省高等学校科学研究优秀青年项目(14B058) 湖南科技大学煤炭资源清洁利用与矿山环境保护湖南省重点实验室开放基金项目(E21701)
【分类号】：TD712.3

【相似文献】