基于ANFIS的煤体瓦斯渗透率预测模型研究

发布时间：2018-05-03 23:34

本文选题：ANFIS + 瓦斯渗透率　；参考：《煤矿开采》2017年01期

【摘要】：为有效预测煤体瓦斯渗透率,预警井下作业时瓦斯浓度变动,利用神经网络的自适应学习能力和模糊推理系统的经验知识建立自适应神经模糊推理系统(ANFIS)预测模型,并基于实验室数据将其预测结果与BP神经网络模型和支持向量机(SVM)模型的预测值作对比。研究结果表明:ANFIS模型的收敛速度快,预测值与实测值相符度高;在误差精度、训练速度和收敛性等方面,其性能优于其他两种模型,可通过有效应力、瓦斯压力、温度和抗压强度对瓦斯渗透率进行高精度的预测。
[Abstract]:In order to effectively predict the gas permeability of coal body and predict the change of gas concentration in underground operation, an adaptive neural fuzzy inference system (ANFIS) prediction model is established by using the adaptive learning ability of neural network and the empirical knowledge of fuzzy inference system. The prediction results are compared with those of BP neural network model and support vector machine (SVM) model based on laboratory data. The research results show that the convergence speed of the 1: ANFIS model is fast, the predicted value is in good agreement with the measured value, and its performance is superior to that of the other two models in terms of error accuracy, training speed and convergence, which can be achieved by effective stress and gas pressure. Temperature and compressive strength are used to predict gas permeability with high accuracy.
【作者单位】：郑州大学管理工程学院;
【基金】：国家自然科学基金资助项目(71271194) 河南省高等学校重点科研项目计划(16A630035) 河南省基础与前沿技术研究计划项目(162300410073) 教育部人文社会科学研究青年基金资助项目(11YJC630291)
【分类号】：TD712

【相似文献】