A Harmonic Level Estimation Method Driven by Accessible Basic Electricity Data

This paper presents a harmonic level estimation methodology for distribution networks based on accessible basic electricity data (A-BED), such as current, voltage, and power. The method employs a long and short-term memory network (LSTM) to assess the harmonic level. Firstly, a chronological law of harmonic emission level is extracted from electricity consumption information data through LSTM, with the aim of fitting the nonlinear relationship between current, voltage and power data and harmonic level. Secondly, the harmonic level of the current system is estimated through real-time A-BED, with the objective of achieving the visualization of harmonic level independently of harmonic monitoring equipment. The case studies illustrate that the proposed method is an efficacious approach for estimating the harmonic level of the current system, based on real-time A-BED from the power grid. The method exhibits high accuracy and stability.