基于BP神经网络优化算法的煤岩识别误差分析
Error Analysis of Coal-rock Identification Based on BP Neural Network Optimization Algorithm
段艳强1
作者信息
- 1. 山西潞安集团潞宁煤业有限责任公司,山西 宁武 036700
- 折叠
摘要
BP(Back Propagation,反向传播)神经网络收敛速度慢,为提高BP神经网络应用水平,采用PSO(Particle Swarm Optimization,粒子群优化)算法优化BP神经网络,提出集 2种算法优势于一体的PSO-BP神经网络算法.通过对BP神经网络优化算法应用原理的了解,结合煤矿项目实例,研究BP神经网络优化算法在煤岩识别中应用的可行性.研究对象为传统BP神经网络和优化后的PSO-BP神经网络,设定误差边界等参数,根据误差曲线、预测结果评价 2 种方法的应用效果.研究表明:PSO-BP神经网络算法的特征值合适,识别精度高,基于BP神经网络优化算法的煤岩识别结果与实际情况具有较高的一致性,采用该优化算法的评价结果可为煤矿生产提供指导.
Abstract
The convergence speed of BP(Back Propagation)neural network is slow.In order to improve the application level of BP neural network,the PSO(Particle Swarm Optimization)algorithm is used to optimize BP neural network,and a PSO-BP neural network algorithm integrating the advantages of two algorithms was proposed.Based on the understanding of the application principle of BP neural network optimization algorithm,the feasibility of applying BP neural network optimization algorithm in coal-rock identification was studied in combination with a coal mine project example.The research object was the traditional BP neural network and the optimized PSO-BP neural network.The error boundary and other parameters were set,and the application effect of the two methods was evaluated according to the error curve and prediction results.The research shows that the characteristic value of PSO-BP neural network algorithm is suitable and the recognition accuracy is high;the results of coal-rock recognition based on BP neural network optimization algorithm have high consistency with the actual situation,and the evaluation results based on this optimization algorithm can provide guidance for coal mine production.
关键词
煤岩识别/BP神经网络/PSO/特征参数/测试分析Key words
coal-rock identification/BP neural network/PSO/characteristic parameter/test analysis引用本文复制引用
出版年
2024
NETL
NSTL
万方数据