摘要
水和煤泥是煤分选过程中的干扰因素,对煤和煤矸石的特征提取、特征可信度和识别精度都有不利影响。基于此,通过数值模拟和实验分析相结合的方法,对煤分选过程中的识别系统及其对水和煤泥因素的适应性进行研究,并提取了煤和煤矸石在水和煤泥作用下的介电和几何特征,形成了特征向量,利用BP神经网络、贝叶斯等分类模型对特征向量的可靠性进行了测试。研究结果表明,含水率和煤泥覆盖对煤矸石介电特性和几何特性的影响可以忽略;模型的识别准确率均超过90%,基于介电特性和几何约束的煤和煤矸石识别方法可以适应多因素的工作条件。
Water and slime are interference factors in the coal sorting process,which have adverse effects on feature extraction,feature reliability,and recognition accuracy of coal and gangue.Based on this,a combination of numerical simulation and experimental analysis was used to study the recognition system and its adaptability to water and slime factors in the coal sorting process.And the dielectric and geometric characteristics of coal and coal gangue under the action of water and slime were extracted,forming feature vectors.The reliability of feature vectors was tested using classification models such as BP neural network and Bayesian.The research results indicate that the influence of moisture content and slime coverage on dielectric and geometric properties of coal gangue can be ignored.The recognition accuracy of model exceeds 90%,and the coal and coal gangue recognition method based on dielectric properties and geometric constraints can adapt to multi factor working conditions.
作者
冉国锋
卢瑞
于继图
Ran Guofeng;Lu Rui;Yu Jitu(Shanxi Shuozhou Pinglu District Longkuang Daheng Coal Industry Co.,Ltd.,Shuozhou 036800,China)
出处
《煤炭科技》
2025年第4期92-96,共5页
Coal Science & Technology Magazine
关键词
煤矸石识别
水
煤泥
介电特性
几何特征
coal gangue recognition
water
slime
dielectric property
geometric characteristic
