The invention discloses a resistance-changing-type single-fault-source diagnosing method for a mine ventilation
system based on
air volume characteristics. A training sample of the fault and
air volume relation is generated through a mine ventilation simulating
system (MVSS), a classification model and a regression model based on a
support vector machine (SVM) are built, and the fault position andequivalent air resistance are diagnosed. The topology structure of a ventilation
system network is complex, change of the air resistance of one of branches of the ventilation system network gives rise to change of the
air volume of the
branch and the ventilation network of a partial area, the air volume serves as a
vector mode so as to distinguish the topology position where a resistance-changing-type fault occurs, and the single-fault-source SVM diagnosing method is put forward to diagnose the resistance-changing-type
single fault source position and fault degree of the mine ventilation system; a final
decision model is built according to the method, the method is simple, human interference is less, connection between data is objectively reflected, and the method is suitable for solvingthe essential non-
linear problem; and the model is determined by a support vector,
data dimension disaster is avoided, and good robustness is achieved.