Covered Overhead Conductor Temperature Coefficient Identification Using A Differential Evolution Optimization Algorithm

The paper describes a method for the temperature coefficient identification for a covered overhead conductor. Numerical and real models of the conductor were used in the paper. The numerical model is made using the Finite Element Method, and contains the material’s temperature characteristics and boundary condition values. The real model presents a part of the covered overhead conductor, on which were mounted thermoelements to measure the temperature in laboratory conditions. The optimization algorithm Differential Evolution (DE) was used to identify the temperature coefficient in order to achieve minimum difference of the temperatures between the two models. The classic algorithm DE is compared to two improved DE algorithms, with dynamic changing of the population size during the optimization process. The two improved DE algorithms are: the DE algorithm with half-reducing of the population, and the DE algorithm with linear-reducing of the population. The results of all three algorithms are compared, and the most favourable choice of algorithm is given for such temperature coefficient calculations.