Design of intelligent fault diagnosis for radar power supply based on SVM

This paper presents a classification model based on SVM for the fault diagnosis of radar array power supplies. A four-class classification framework was developed by collecting three key parameters: output current, output voltage, and module temperature of the array power supply, encompassing the normal state and three distinct fault conditions. The one-versus-one strategy was employed to decompose the four-class classification problem into multiple binary classification tasks. Experimental results demonstrate that this approach achieves high accuracy in fault diagnosis, offering effective technical support for the health status assessment of the radar small array system power supplies.