The purpose of this paper is to analyze two probabilistic models of atmospheric icing on overhead transmission line conductors, concerning the combined wind-on-ice loads and the specification of radial ice thickness v/s linear ice weight. These models are important for design purposes and greatly influence the mechanical behavior and reliability of the lines. Meteorological data from 150 icing measurement sites in Quebec was used for Monte Carlo simulations with various distributions for wind speed and icing. Statistical analysis indicates some important deficiencies of the icing models adopted. A method for determining the combined wind-on-ice loads is proposed, using the largest extreme (type I) distribution for icing and the Weibull distribution for wind speed. The reduction factor for wind speed associated with ice events was found to be in the range of 0.4 to 0.6. The specification of ice linear weight on conductors instead of ice radial thickness is proposed for probability based line design.
Overhead power lines in service, their mechanical behavior and structural reliability, are subjected to various meteorological loads such as wind and icing. The random character of these loads requires statistical approach when their magnitudes are determined for design purposes. Despite many years of research efforts, problems related to the probabilistic aspects of traditional ice accretion model still persist. Two such deficiencies are:
correct prediction of the simultaneous occurrence of extreme ice and wind and
the difference between the specified ice radial thickness instead of ice linear weight.
It is almost impossible to enumerate all the publications in this area, which are presented in general as guidelines and recommendations by national codes (Ghannoum, 1993), or by institutions like the International Electrical Commission (IEC, 1991) and the American Society of Civil Engineers (ASCE, 1984). By way of illustration, the normalized meteorological conditions in Quebec are summarized in Table 1.
