基于再生核RBF神经网络的瓦斯突出预测系统  

Gas Outburst Prediction System Based on Reproducing Kernel RBF Neural Network

作　　者：彭泓[1,2] 王聪[1] 王斌[2] 

机构地区：[1]中国矿业大学机电学院,北京100000 [2]辽宁工程技术大学电气与控制工程学院,辽宁葫芦岛125105

出　　处：《计算机测量与控制》2012年第6期1463-1466,共4页Computer Measurement &Control

基　　金：国家自然科学基金项目(50874059)

摘　　要：该设计的瓦斯突出预测系统由数据采集,数据传输和数据处理三部分组成;首先使用层次分析法和MATLAB选择出了瓦斯突出影响因素,然后使用TMS320C6713和PCI总线技术设计了数据采集和传输系统,同时采用再生核算法来进行RBF神经网络的训练,通过W12[a,b]空间插值逼近的方法,把RBF神经网络的训练转换为解线性方程组,最后使用LABVIEW,MATLAB和CCS混合编程实现了再生核RBF神经网络的训练和仿真以及TMS320C6713软件开发,准确地预测出了瓦斯突出。Gas outburst prediction system consists of three parts, which is data collection, data transmission and data processing, de- signed in the paper. First, the factors of affecting gas outburst was Selected using the AHP and MATLAB, Then, data collection and trans- mission system is designed with TMS320C6713 and the PCI bus technology, At the same time, RBF neural network is trained, using Repro- ducing Kernel algorithm, By the space W~ Ea, b] interpolation approximation, training of the RBF neural network is converted into seeking the solution of the linear equations system, Finally, Reproducing Kernel RBF Neural Network training and simulation and TMS320C6713 Software Development is achieved, using the Mixed programming of LABVIEW, MATLAB and CCS, outburst was accurately predicted.

关 键 词：再生核RBF神经网络 瓦斯突出 PCI总线 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]