基于LVQ-CPSO-BP算法的煤体瓦斯渗透率预测方法研究

发布时间：2018-05-13 23:14

本文选题：瓦斯渗透率 + 学习向量量化神经网络(LVQ)　；参考：《采矿与安全工程学报》2017年02期

【摘要】：针对BP神经网络算法对煤体瓦斯渗透率预测精度低问题,筛选出影响预测精度的5个主要因素——1个宏观因素(煤层埋深)和4个微观因素(有效应力、温度、瓦斯压力、抗压强度),提出一种基于学习向量量化神经网络(LVQ)分类、混沌粒子群算法(CPSO)优化、BP神经网络预测的LVQ-CPSO-BP煤体瓦斯渗透率预测方法。从宏观上确定临界值将煤层埋深划分为2层;基于有效应力与瓦斯渗透率之间存在拐点关系,从微观上确定拐点值将有效应力划分为2段;采用LVQ将4个微观样本参数依据拐点特征进行分类识别,采用BP神经网络进行学习训练并输出预测结果,并用CPSO对BP神经网络的权值和阈值进行优化;基于样本案例对本文构建的LVQ-CPSO-BP算法进行预测结果验证,并与BP算法、GA-BP算法及PSO-BP算法预测的结果进行对比分析。结果表明:LVQ分类正确识别率较高,CPSO-BP算法预测精度较好,且优于其他3种算法。LVQ-CPSO-BP算法总体预测值与实测值吻合度高,尤其当有效应力减小时,预测精度更高。
[Abstract]:In view of the low accuracy of BP neural network algorithm in predicting coal gas permeability, five main factors affecting the prediction accuracy are selected: one macroscopic factor (coal seam depth) and four microscopic factors (effective stress, temperature, gas pressure). Based on learning vector quantization neural network (LVQ) classification and chaotic particle swarm optimization (PSO), a LVQ-CPSO-BP coal gas permeability prediction method based on BP neural network is proposed. According to the inflection point relationship between effective stress and gas permeability, the effective stress is divided into 2 sections from the microscopic point of view, and the critical value is determined macroscopically, and the buried depth of coal seam is divided into two layers, based on the inflection point relationship between effective stress and gas permeability. The four microscopic sample parameters are classified and identified according to the inflection point feature by LVQ, the BP neural network is used for learning training and the prediction results are outputted, and the weights and thresholds of the BP neural network are optimized by CPSO. The prediction results of the LVQ-CPSO-BP algorithm constructed in this paper are verified by a sample case, and compared with the prediction results of the BP algorithm GA-BP algorithm and the PSO-BP algorithm. The results show that the accuracy of the CPSO-BP algorithm is higher than that of the other three algorithms. The overall prediction value of LVQ-CPSO-BP algorithm is in good agreement with the measured value, especially when the effective stress is reduced, the prediction accuracy is higher.
【作者单位】：新疆大学电气工程学院;清华大学自动化系;新疆煤炭设计研究院有限责任公司矿山所;
【基金】：国家自然科学基金项目(51264036) 新疆自然科学基金项目(2015211C271)
【分类号】：TD712.3

【相似文献】