Abstract：A fault diagnosis model for isolation switches based on Markov transition field（MTF），catch fish

optimization algorithm（CFOA）improved pyramid vision transformer（PVT）was proposed to address the

problems of complex fault situations，strong noise in signal extraction，and insufficient feature extraction. Firstly，

the MTF algorithm was used to convert the one-dimensional vibration signal of the time sequence into a twodimensional image，preserving the relevant temporal characteristics in the one-dimensional signal. Secondly，the obtained two-dimensional image was input into the PVT model optimized by the CFOA for extracting features from the image. Finally，the obtained model was applied to the diagnosis of isolation switch faults. The experimental results show that under different fault conditions，this method can achieve a fault classification accuracy of 98.25% for isolation switches，demonstrating the superiority of this method.

      Key words：Markov transition field（MTF）；pyramid vision transformer（PVT）；isolating switch；fault diagnosis