CPC G06N 3/047 (2023.01) [G06F 17/14 (2013.01); G06N 3/08 (2013.01)] | 8 Claims |
1. A transformer failure diagnosis method based on an integrated deep belief network, comprising:
obtaining a plurality of vibration signals of transformers of various types exhibiting different failure types, retrieving a feature of each of the vibration signals, and establishing training data through the feature retrieved corresponding to each of the vibration signals;
training a plurality of deep belief networks exhibiting different learning rates through the training data and obtaining a failure diagnosis correct rate of each of the deep belief networks; and
keeping target deep belief networks corresponding to the failure diagnosis correct rates that satisfy requirements, building the integrated deep belief network through each of the target deep belief networks, further comprising:
obtaining a mean correct rate Mean diagnosis accuracy of N of the failure diagnosis correct rates, N is a number of the deep belief networks;
eliminating the deep belief networks corresponding to the failure diagnosis correct rates lower than the mean correct rate Mean diagnosis accuracy and obtaining the remaining target deep belief networks;
obtaining an extra correct rate Extra accuracyt of a tth target deep belief network through Extra accuracyt=Diagnosis accuracyt−Mean diagnosis accuracy, wherein t=1 . . . T, T is a number of the target deep belief networks, and Diagnosis accuracyt is the failure diagnosis correct rate of the tth target deep belief network;
distributing a weight value Weightt to the tth target deep belief network through
and
forming the integrated deep belief network through the target deep belief networks together with weight values corresponding thereto,
and performing a failure diagnosis on the transformers through the integrated deep belief network.
|