Suppression of Strong Cultural Noise in Magnetotelluric Signals Using Particle Swarm Optimization-Optimized Variational Mode Decomposition

To effectively separate strong cultural noise in Magnetotelluric (MT) signals under strong interference conditions and restore the true forms of apparent resistivity and phase curves, this paper proposes an improved method for suppressing strong cultural noise based on Particle Swarm Optimization (PSO) and Variational Mode Decomposition (VMD). First, the effects of two initial parameters, the decomposition scale K and penalty factor alpha, on the performance of variational mode decomposition are studied. Subsequently, using the PSO algorithm, the optimal combination of influential parameters in the VMD is determined. This optimal parameter set is applied to decompose electromagnetic signals, and Intrinsic Mode Functions (IMFs) are selected for signal reconstruction based on correlation coefficients, resulting in denoised electromagnetic signals. The simulation results show that, compared to traditional algorithms such as Empirical Mode Decomposition (EMD), Intrinsic Time Decomposition (ITD), and VMD, the Normalized Cross-Correlation (NCC) and signal-to-noise ratio (SNR) of the PSO-optimized VMD method for suppressing strong cultural noise increased by 0.024, 0.035, 0.019, and 2.225, 2.446, 1.964, respectively. The processing of field data confirms that this method effectively suppresses strong cultural noise in strongly interfering environments, leading to significant improvements in the apparent resistivity and phase curve data, thereby enhancing the authenticity and reliability of underground electrical structure interpretations.