Online-Vorhersageverfahren und -Vorrichtung zur Messung des Fehlers eines Elektronischen Stromwandlers

Details

Veröffentlichungs-Nr.

EP4790444

Anmeldetag

20. November 2025

Veröffentlichung

25. Juni 2026

Rechtsraum

EP

IPC

G01R35/02

Abstract

The present invention provides a method and an apparatus for online prediction of a measurement error of an electronic current transformer. The method includes: obtaining environmental data of an environment where the electronic current transformer is located, and obtaining error data of the electronic current transformer in an automatic prediction manner by a neural network, thereby implementing online automatic prediction of an error of the electronic current transformer, improving efficiency for determining the error of the electronic current transformer, and further enhancing calibration efficiency of the electronic current transformer. Further, a target swimming penetration neural network includes a retention layer, and the retention layer is used to store an output from a hidden layer in a previous iteration. Compared with a conventional neural network, a structure of this retention layer can effectively enhance both the overall and local stability of the swimming penetration neural network, thereby improving accuracy of error prediction by the electronic current transformer.

Anmelder

Firma

State Grid Zhejiang Tongxiang Electric Power Supply Company

Land

🇨🇳 China

Fachgebiete