Traditional state assessment methods often rely on single frequency signals and can only detect faults within a specific frequency range, making it difficult to fully reflect the overall state of the equipment. This article introduces multi-frequency signal processing technology to evaluate the status of GIS (Geographic Information System) equipment in substations. Firstly, the signal’s global spectrum features are extracted using the Fourier transform, and its properties at various time and frequency scales are obtained by time-frequency analysis using the wavelet transform. Then, spectral and time-frequency feature parameters are extracted from the decomposed signal, and decision level fusion using voting method is used to fuse the signal features of different frequencies and types, and output a comprehensive feature vector. Finally, the LSTM (Long Short-Term Memory) algorithm is used to analyze the fused feature vectors and evaluate the status of GIS equipment. The experiment was based on the log data of a substation company in Wuhan from June to December 2023, and combined with multi-frequency signal processing to evaluate the status of GIS equipment. The results showed that the accuracy of state evaluation by integrating multi-frequency signal processing and LSTM method reached 98.25%. Compared to the accuracy based on a specific frequency range, it has improved by 8.53%, and the shortest response time is only 1.8s. Experiments have shown that multi-frequency signal processing plays an important role in the comprehensive evaluation of GIS equipment status in substations, greatly improving the accuracy of evaluation, achieving accurate reflection of the overall status of GIS equipment, and promoting the stable operation of substations.
