Research on Self-Detection Method of Current Transformer Measurement Error Based on Data Secondary Correction

The current transformer is a critical sensing device for real-time monitoring of power system operations. However, its performance can degrade over time, leading to measurement errors that compromise accurate detection of primary signals. Traditional error detection methods rely on periodic comparisons with standard current transformers under live or offline conditions, which lack real-time responsiveness. To address this limitation, this article proposes an online self-detection method for current transformer measurement errors based on secondary data correction. The proposed method consists of three main steps: 1) using affine transformation to correct synchronous three-phase current data from nodes, establishing a new correlation-constrained comparison standard; 2) extracting residual components containing error information via the empirical wavelet transform (EWT)-autoregressive moving average (MA) method, followed by a secondary correction to enhance error features; and 3) dynamically evaluating the corrected data using an improved moving window principal component analysis (PCA) approach. This framework enables long-term, real-time error self-detection. Experimental results from field data demonstrate that the method achieves a minimum detectable error that meets the 0.2 accuracy requirement.