An analysis of the icing of test cables

Abstract

In cold countries, the design of transmission lines and communication networks requires the knowledge of ice loads on conductors. Atmospheric icing is a stochastic phenomenon and therefore probabilistic design is more and more used for structure icing analysis. For strength and reliability assessments, a data base on atmospheric icing is needed to characterize the distributions of ice load and corresponding meteorological parameters. A test site, located on Mt. Valin, near Chicoutimi, Quebec, Canada, where icing is frequent, has been used to obtain field data on atmospheric icing. The experimental installation is composed of instrumented but non-energized test cables, meteorological instruments, a data acquisition system, and a video recorder. Rime, glaze, wet snow, and mixtures of these types of ice can produce large ice accretions which can be dangerous for land-based structures. These different icing events must be distinguished, before statistical analysis, in the data base on atmospheric icing. This has been done by comparison of data from a precipitation gauge, an icing rate meter, and a temperature sensor. An analysis of selected icing periods recorded during the 1992-1993 winter season is performed. Curves representing the ice load variation with time on four different conductor types are presented, together with the corresponding variations of air temperature, icing rate meter signal, precipitation gauge signal, and wind velocity component normal to the main test line. The analysis of data shows a typical evolution of icing events with air temperature variation during the winter season. The first two events, in October 1992, are accretions due to precipitation icing. The next three icing periods, from the end of October to the end of November 1992, are composed of successive accretions of rime and glaze. Then, during the coldest period of winter, between the end of November 1992 and March 1993, the five icing periods observed are mainly rime accretions due to in-cloud icing, with only very small precipitation. The next icing period, in April 1993, is composed of three rime accretions and one icing event caused by a mixture of in-cloud icing and precipitation icing. These latter four icing events are separated by partial or total shedding of the ice induced by melting. The last icing period, in May, is a mixture of in-cloud icing and precipitation icing, followed by shedding of ice which is also induced by melting. A comparative analysis of the ice load on the four test cables is drawn from the data, and typical accretion and shedding parameters are calculated separately for icing events related to in-cloud icing and precipitation icing.