Partial discharge measurement is an effective method to reflect the insulation defects of electrical equipment.Detection of partial discharge by radio frequency can achieve non-intrusive measurements and has good application prospects.By constructing finite element model of multi-gap loop(MGL)antenna sensor the factors affecting the sensor performance are analyzed.Combined simulations with the multi-population genetic algorithm(MPGA)and the finite element method(FEM)are carried out to optimize the structural parameters of the MG

L sensor aiming at the requirement of radio frequency detection of partial discharge in cable intermediate joint.The sensitivity of the optimized MGL sensor is determined through FEM model of gigahertz transverse electromagnetic(GTEM)cell.The results show that the center frequency of the sensor is 150MHz

with a

S

11

of less than -10dB in the frequency band of 25~590MHz

and an average sensitivity of 7569V·(A/mm)

-1

meeting the industry requirement of sensitivity of radio frequency sensors for partial discharge detection.The research achievement can contribute to the realization of a novel radio frequency sensor for non-intrusive detection of partial discharge in cable intermediate joint.