煤矿膏体充填测控系统与RBF神经网络建模

通过对充填实验工艺的研究,设计了可靠的充填实验过程计算机测控系统,实现对试验过程的监测与控制。煤矿现场井下管路的出口压力是主要的控制参量,它必须满足变化的射程。而现场井下环境极其恶劣,难以测量。采用RBF神经网络建模方法,获得了膏体容重、塌落度、质量浓度、平均流速与出口压力值之间的多变量非线性函数模型。经模拟实验系统检验,模型预报的出口压力满足精度要求,且具有快速性和实时性的特点,解决了现场实测困难的问题,为现场工程应用推广打好了基础。

Paste-filling Measurement Control System in Mining and RBF Neural Net Modeling

By studying the paste-filling technology, a reliable computer measurement control system is designed to realize the automatic measurement and control in the process of filling experiment. The exit pressure in filling field is a main parameter, which must meet the filling range. However, the circumstance of filling field is so poor that it is difficult to be measured. RBF net is used to build the model. A multivari-able non-linear function model between paste-volumeweight, subside, quality concentration, average flow rate and exit pressure can be obtained. After simulation experiments, the predictive exit pressure can meet precision demand, the method also has rapid and real-time characteristic, so it can solve the difficult measure question in mine filling. The method lay a good foundation for field application.