The Malaysian power industry was a regulated, organized electricity supply wise inherited from British colonization and showed to be robust and reliable throughout the millennium. The electricity sector is always committed to ensuring that supply reliability is preserved on a continual basis without any major disturbance. The most prevalent downtime causes are closely linked to high voltage (HV) equipment malfunctions or defects, especially transformers, switchgears and cables. The consequences of these occurrences are catastrophic and lead to the loss of millions of dollars in terms of the refurbishment or replacement of HV infrastructure. Power utility company has mostly attempted to stop this event from happening through HV equipment monitoring to classify the irregularity before it is more severe. Partial discharge (PD) event is among the phenomena that is being measured and evaluated. PD is commonly found in solid, gaseous or fluid form correlated with void. In this research, an unsupervised approach to the neural network is recommended for PD classification. This journal introduces the PD classification use an analysis based on the combo of self-organizing maps (SOMs) and correlation analysis. This project established that PD data was successfully correlated and clustered using MATLAB software's SOM Toolbox and correlation tool to identify the type of PD in a HV apparatus.