矿井安全监控平台中的智能决策系统研究

Research on intelligent decision-making system of mine safety monitoring platform

针对当前矿井监控系统只提供阈值报警的不足,充分利用监控系统数据库中丰富的历史数据,先甄选出正常数据样本与异常数据样本,分别对正常、异常样本建模,再提取各样本的特征向量,最后利用支持向量机训练各样本的特征向量,得到以矿井瓦斯为主的在线预测模型。仿真结果表明,支持向量机预测模型能够实现矿井瓦斯的在线测量与危险性提前预测,为矿井管理人员提供智能决策指导,进一步完善了矿井监控系统的整体功能。

In view of the shortcoming of the current mine monitoring system only providing the threshold alarm,this paper makes full use of the rich historical data in the database of monitoring system,selects the normal data samples and the abnormal data samples first,respectively models the normal and abnormal samples and then extracts the feature vector of each sample,and finally using the support vector machine to train the feature vector of each sample to get the on-line forecasting model mainly based on mine gas.The simulation results show that support vector machine prediction model can realize the on-line measurement and risk prediction mine gas.It provides intelligent decision guidance for mine managers,and further improves the overall function of mine monitoring.