Performance Evaluation Framework for Broadband Measurement Methods in Power Systems Based on Deep Learning

With the rapid integration of distributed energy resources and power electronic devices in modern power systems, broadband harmonic and interharmonic phenomena have become critical challenges, affecting grid stability and power quality. Accurate measurement of wideband signals is essential, and equally important is the performance evaluation of measurement methods. This paper proposes a deep learning-based performance evaluation framework for wideband measurement methods in power systems, designed for the automation and quantitative assessment of broadband measurement methods in power systems. The framework utilizes a deep neural network (DNN) model to analyze measurement errors in the measured signals, including frequency, amplitude, and phase deviations. A standardized dataset was generated, with Gaussian noise added to simulate real-world measurement errors. Experimental results demonstrate the effectiveness of the model. Three simulated measurement methods were evaluated, and the evaluation results validate the model’s effectiveness and its sensitivity to noise and bias. The framework introduces a novel automated evaluation approach, overcoming the limitations of traditional manual assessments.