In order to solve the problem of insufficient research on overheating and optimization design for terminal connectors in UHV DC converter station

the temperature rise of the terminal connectors are studied by the finite element method

and the reliability of the terminal connectors is calculated by combining the radial basis function（RBF） neural network with the Monte-Carlo method. A terminal connector’s optimal mathematical model constrained by its reliability is established

and the optimum design method based on this model is given. The results show that

under the condition of high current

the temperature at the overlapping parts of the terminal connectors is higher than that of other parts

and increases exponentially with time. At the beginning of current loading

the heat flowing into the terminal connectors is greater than the heat flowing out

and the temperature rise rate of the overlapping parts is high. With the increase of current loading time

the heat flowing into the connector terminals increases gradually and approaches the heat flowing out

until the temperature at the overlapping parts no longer increases since the terminal connectors reaching a state of heat balance. The temperature at the overlapping parts of the terminal connectors

with decrease in cross-section size

increases nonlinearly

and the influence of cross-section thickness on the temperature rise of the overlapping parts is significantly higher than that of cross-section width. When the optimum design meets the reliability requirements by using the optimization mathematical model

the weight of the terminal connectors can be reduced by about 15% and the operational reliability can be improved.