基于传感数据的矿井坍塌风险预测模型仿真

Simulation of Mine Collapse Risk Prediction Model Based on Sensor Data

针对矿井施工区域内的执行进展情况复杂,全局控制困难的问题,提出一种基于传感数据的矿井坍塌风险预测模型。按照Hadoop传感平台的连接需求,通过插入数据浏览表的方式,完成基于传感技术的矿井坍塌风险数据处理。在此基础上,迎合风险数据的过拟合处理原则,实施归一化的预测统计操作,最后延续核函数确定理论,实现矿井坍塌风险预测模型构建。仿真结果表明,与支持向量机风险预测模型相比,新型风险预测模型所具备的数据完整性大幅增强,在传输敏感性方面,与动态博弈监管模型与Map模型相比,也出现明显的数值水平上升趋势,坍塌风险预测误差较低。

In this paper, a model to predict mine collapse risk based on sensing data was proposed. According to the connection requirements of Hadoop sensing platform, the mine collapse risk data based on sensing technology was processed by inserting data browse table. On this basis of overfitting principle of risk data, we normalized the statistical prediction. Finally, we continued to determine the theory of kernel function. Thus, we built the mine collapse risk prediction model. Simulation results show that, compared with the risk prediction model of support vector machine, the data integrity of new risk prediction model is greatly enhanced. In terms of transmission sensitivity, the new model also has an obvious rising trend of numerical level compared with the dynamic game regulation model and Map model. Meanwhile, the error of collapse risk prediction is low.