A New Data Fusion Approach for a Log Amplifier Channel

摘要

A digital radio frequency monitor is an on-line monitoring system for stator windings of generators. A log amplifier channel is an important unit of digital radio frequency monitor. The characteristic of the channel is no ideal. In order to improve the stability of the log amplifier channel, this paper presented a new data fusion approach based on neural network. The neural network applied to characteristic correction of the radio frequency current transformer and the receiving channel. This paper describes the effects of crystal filter characteristics on the setting values of alarm level and compensation. The data fusion network structure and learning algorithm are also discussed. The new data fusion structure is based on neural network. The fusion network consists of 4 components. The data fusion approach presented might make RF level measurements easier and more practical. A practical example is presented to show that the proposed network and data fusion method are very feasible and efficient.