CAJ | 학술논문

为解决井下风速传感器获得的风速异常数据确定故障源位置问题，采用RBF神经网络方法确定可能引起各分支风速异常的分支集合，即建立通风系统故障巷道范围库，再通过对分支的灵敏度进行排序来选择故障巷道诊断的优先级。研究结果表明：RBF 神经网络被训练好后，就可以不用建立具体的数学模型，得到整个网络各分支风量与风阻之间的关系。
위해결정하풍속전감기획득적풍속이상수거학정고장원위치문제，채용RBF신경망락방법학정가능인기각분지풍속이상적분지집합，즉건립통풍계통고장항도범위고，재통과대분지적령민도진행배서래선택고장항도진단적우선급。연구결과표명：RBF 신경망락피훈련호후，취가이불용건립구체적수학모형，득도정개망락각분지풍량여풍조지간적관계。
In order to solve the problem of getting abnormal air velocity data by using air velocity transducer to determine the fault source location, the branch collection that may cause the abnormal air velocity is determined by using the RBF neural network method, and the ventilation system fault roadways scope library is established. Then the priority level of fault roadway diagnosis is selected through sorting branch sensitivity. The results show that specific mathematical models do not need to be established by using the trained RBF neural network, which will be able to get the relationship between roadway air volume changes and lead to its changes each branch drag.