Fault Classification Method for Electric Power IoT Equipment Based on Edge Computing
Abstract
With the advent of the era of big data, a large amount of data generated by local sensors in the power Internet has greatly increased the calculation amount of server terminals and increased their processing burden. Traditional centralized fault diagnosis systems have been unable to meet the needs of real-time diagnosis, and it is urgent to offload calculations. This paper proposes a fault detection and classification method suitable for electric power IoT devices based on edge computing technology. This method uses DNN neural network to implement fault detection at the edge layer, find out all fault data, and then upload the fault data only to the cloud center layer. At the cloud center layer, HTM neural network is used to classify the fault data to complete the fault diagnosis. Finally, experiments were performed using laboratory data from Case Western Reserve University in the United States to prove the reliability of the algorithm.
Keywords
edge computing, fault detection, DNN, HTM
DOI
10.12783/dtetr/mcaee2020/35006
10.12783/dtetr/mcaee2020/35006
Refbacks
- There are currently no refbacks.