The imbalanced transformer fault samples and the tendency of fault diagnosis models to fall into local optima may degrade fault identification performance. To address the issues above

this paper proposes a transformer fault diagnosis method that integrates a Wasserstein generative adversarial network(WGAN) with a support vector machine(SVM) optimized by an improved sparrow search algorithm(ISSA). First

a WGAN-based data augmentation strategy is employed to expand the sample set and mitigate class imbalance. Second

Logistic chaotic mapping

a random walk strategy

and an adaptive firefly algorithm are introduced to improve the standard sparrow search algorithm(SSA)

thereby enhancing its parameter optimization capability. The improved algorithm is then used to optimize the key parameters of SVM

and an ISSA-SVM fault diagnosis model is established. Finally

fault diagnosis experiments are conducted on the WGAN-augmented dataset. The experimental results show that the proposed method maintains high diagnostic performance under imbalanced sample conditions

achieving a diagnostic accuracy of 96.60%

outperforming the SSA-SVM model and other data balancing methods.