EKF-ENN算法在瓦斯涌出量预测中的应用
发布时间：2018-07-17 07:07
【摘要】：针对瓦斯涌出量的多影响因素预测问题,提出将扩展卡尔曼滤波算法与Elman神经网络有机结合并应用于瓦斯涌出非线性系统的动态辨识.带有整定因子的EKF滤波器通过实时对网络的权值、阈值进行全局寻优,建立基于EKF-ENN耦合算法的绝对瓦斯涌出量预测模型,利用矿井监测到的各项历史数据进行预测试验.研究结果表明:该模型的预测平均相对误差为1.67%;平均相对变动值ARV为0.000 768 1.EKF优化后的Elman神经网络预测模型相比于其他预测模型,具备更高的预测精度与更强的泛化能力.
[Abstract]:Aiming at the problem of multi-factor prediction of gas emission, the extended Kalman filter algorithm and Elman neural network are organically combined and applied to the dynamic identification of gas emission nonlinear system. The EKF filter with tuning factor is used to optimize the weights and thresholds of the network in real time. Based on the EKF-ENN coupling algorithm, the prediction model of absolute gas emission is established, and the prediction experiments are carried out by using the historical data monitored by the mine. The results show that the prediction average relative error of the model is 1.67 and the average relative variation value ARV is 0.000. EKF optimized Elman neural network model has higher prediction accuracy and stronger generalization ability than other prediction models.
【作者单位】： 辽宁工程技术大学电气与控制工程学院;新疆师范大学外国语学院;
【基金】：国家自然科学基金项目(5127411) 辽宁省教育厅基金项目(L2012119)
【分类号】：TD712.53
[Abstract]:Aiming at the problem of multi-factor prediction of gas emission, the extended Kalman filter algorithm and Elman neural network are organically combined and applied to the dynamic identification of gas emission nonlinear system. The EKF filter with tuning factor is used to optimize the weights and thresholds of the network in real time. Based on the EKF-ENN coupling algorithm, the prediction model of absolute gas emission is established, and the prediction experiments are carried out by using the historical data monitored by the mine. The results show that the prediction average relative error of the model is 1.67 and the average relative variation value ARV is 0.000. EKF optimized Elman neural network model has higher prediction accuracy and stronger generalization ability than other prediction models.
【作者单位】： 辽宁工程技术大学电气与控制工程学院;新疆师范大学外国语学院;
【基金】：国家自然科学基金项目(5127411) 辽宁省教育厅基金项目(L2012119)
【分类号】：TD712.53
【参考文献】
相关期刊论文 前10条
1 邵良杉;付贵祥;;基于数据挖掘的瓦斯信息识别与决策[J];辽宁工程技术大学学报(自然科学版);2008年02期
2 邵良杉;芦春霞;;不同井型煤炭项目环境效益的生态折现率仿真[J];辽宁工程技术大学学报(自然科学版);2012年03期
3 邵良杉;付贵祥;;基于数据融合理论的煤矿瓦斯动态预测技术[J];煤炭学报;2008年05期
4 王晓路;刘健;卢建军;;基于虚拟状态变量的卡尔曼滤波瓦斯涌出量预测[J];煤炭学报;2011年01期
5 邵良杉;张宇;;基于小波理论的支持向量机瓦斯涌出量的预测[J];煤炭学报;2011年S1期
6 吕伏;梁冰;孙维吉;王岩;;基于主成分回归分析法的回采工作面瓦斯涌出量预测[J];煤炭学报;2012年01期
7 付华;姜伟;单欣欣;;基于耦合算法的煤矿瓦斯涌出量预测模型研究[J];煤炭学报;2012年04期
8 张翔;王佰顺;徐硕;杨培陪;;改进的神经网络在瓦斯涌出量预测中的应用[J];煤炭技术;2013年01期
9 魏春荣;李艳霞;孙建华;米红伟;李s，
本文编号：2129531
本文链接：https://www.wllwen.com/kejilunwen/anquangongcheng/2129531.html
