Background: In this paper, as to the nonlinearity of the Linear Power Amplifier (LPA) in
Protective Relay Test Device (PRTD), a digital predistorter based on the Adaptive Neuro-Fuzzy
Inference System (ANFIS) is proposed and is named as ANFIS digital predistorter.
Method: By merely modifying the software of the Industrial Personal Computer (IPC) of the PRTD,
ANFIS digital predistorter can upgrade the traditional PRTD to a new generation one, making the
PRTD fulfill the accuracy requirement of fault waveform reproduction and the limited bandwidth and
output capacity of the LPA be better exploited. The ANFIS digital predistorter adaptively learns the
total nonlinearity of both LPA and its load, and therefore the nonlinear error resulting from any
component can all be compensated, which lowers the accuracy requirement of every single component.
Conclusion: And with the quasi-online learning mode, the ANFIS digital predistorter can be realized on
general IPCs, and thus maintains the high performance-price ratio of the traditional PRTD.