Taking a large single-phase generator transformer product as the research object, this paper applied the finite element simulation method to analyze the vibration characteristics of its core. Firstly, through the analysis of vibration theory, the vibration laws and characteristics of the core are clarified, and a three-dimensional equivalent model of the transformer is constructed. The B-H curve of the silicon steel sheet is measured through experiments and used for the assignment of the core material when calculating the electromagnetic field later. Then, based on the simulation calculation of multi-physical field coupling, the no-load current on the primary side, the distribution and variation characteristics of the magnetic field inside the core are solved and analyzed. On this basis, the sequential coupling method was adopted to solve the displacement distribution of the core vibration and the displacement changes at different position points and conduct a comparative analysis. Subsequently, the accuracy of the simulation calculation method was verified through the test of a small prototype. Finally, based on the comparison of the dry and wet modal simulation results, the impact of transformer oil on the vibration characteristics of the core was evaluated and analyzed. It can be seen from the analysis that the core vibration is generally more intense at the upper part and corners; the impact of the internal oil flow on the vibration of the body of large transformers is relatively complex and thus cannot be ignored.

To address the problems of low diagnostic accuracy and weak robustness in bearing fault diagnosis, this paper establishes a theoretical framework for frequency-domain-enhanced residual network diagnosis driven by a multi-stage adaptive snow goose algorithm (SGA), breaking through the fundamental bottleneck of structural rigidity in traditional methods. The dual-flexible residual network achieves adaptive matching between network topology and fault characteristics through variable convolutional layers and dynamic skip connections. The frequency-domain attention module, based on Fourier transform and magnitude-phase separation techniques, provides in-depth analysis of the frequency-domain discrimination mechanism in vibration signals. The improved snow goose algorithm integrates multi-stage adaptation, opposition-based learning, elite memory management strategies, and cooperative foraging mechanisms, enhancing the optimization performance of SGA. Validation on the Jiangnan University and Case Western Reserve University (CWRU) bearing fault datasets demonstrates that the proposed method achieves diagnostic accuracies of 99.81% and 99.78%, respectively, representing improvements of 4.62%, 6.41%, and 8.76% over mainstream methods such as ResNet, WDCNN, and Transformer. Furthermore, when faced with changes in data structure, the proposed method exhibits a performance degradation of only 0.43%, which is significantly better than other methods, demonstrating excellent robustness and cross-domain generalization capability, and it has certain practical significance.