To address the complex evolution of partial discharge within the inter-turn insulation of trans-former windings

and the limitations where single detection signals are susceptible to interference and struggle to accurately characterize stage-specific features

a comprehensive multi-physical quantity joint detection platform was constructed. During the experimental process

pulsed current

ultra-high frequen-cy

and ultrasonic signals were concurrently measured

and their respective evolution patterns throughout the discharge progression were rigorously analyzed. By utilizing the relative relationships of signal energy

variations in pulse repetition frequency

and spectral characteristics as key feature parameters

the partial discharge development process was categorized into three sequential stages: initiation

development

and pre-breakdown. Furthermore

the underlying physical mechanisms governing each discharge stage were inferred based on the multi-dimensional data. This multi-parameter approach effectively overcomes the limitations of traditional single detection methods

which are prone to interference and misjudgment. It significantly enhances the reliability and engineering applicability of state assessment

providing a robust analytical basis for transformer insulation fault diagnosis.