Modern meteorology is used to better characterize overhead line loading events.
To meet the constantly increasing energy requirements of developing societies as well as industrialized countries, electrical infrastructure has to expand into new land and mountainous terrain. In these locations, however, there is often limited knowledge about the frequency or magnitude of adverse weather phenomena, which are important considerations for utilities in the design of overhead lines. Utilities also must consider the maintenance and operation of these systems in addition to the contingency planning that is necessary to prepare for such extreme weather conditions.
The evaluation of design loads attributable to wind and ice, as well as the operational and operating characteristics such as conductor galloping and fatigue experienced by overhead line conductors, is dependent on local weather conditions. This is especially the case for overhead lines erected in areas that have a complex topography where it is almost impossible to obtain the necessary weather details with a spatial resolution adequate for the route of a transmission line, from either general weather observation data or dedicated measurements taken for limited time periods.
Over the last few decades, tremendous developments have been made in global weather observations and computer capacities, so the knowledge of physical and dynamical processes in the atmosphere has progressed similarly. As a result, the quality and reliability of modern weather forecasts have improved significantly.
Meteorology Information Systems
It is now possible to describe, in detail, the water cycle and related phase transitions in clouds as well as the formation of precipitation. Additionally, the accuracy of forecasts of air temperature, wind speed and wind direction have improved significantly. However, to obtain such details in the lower atmosphere, it is necessary to include adequate details of the land and sea surface properties, such as topography, land surface conditions (forests, towns, lakes, farm land and snow cover) and sea surface temperatures. By using nesting technology from global scales, it is possible to model the local weather down to spatial scales relevant for the span lengths of overhead power lines in 3-D topography.
Every six hours, the state of the atmosphere is analyzed on a global scale and all parameters are stored in a 3-D grid database, covering the globe and throughout the atmosphere. This database represents a synthesis of all measurements and observations from regular weather stations, automatic stations in remote areas of the earth and radio soundings, in addition to data from radars and satellites. Each grid point provides comprehensive and reliable information on the state of the atmosphere in such detail that hardly any single weather station could equal it, unless it was a station at a specific site.
Weather research and forecasting (WRF) is the model frequently used for advanced atmospheric applications. It is a state-of-the-art mesoscale numerical weather-prediction system used both for operational weather forecasting and in atmospheric research. WRF solves complex equations for all the important physical processes in the atmosphere, such as winds, temperatures, stability, liquid-water content in clouds, and types and amounts of precipitation, based on initial fields of assimilated weather observations.
However, the WRF model is not able to calculate global weather with a satisfactory level of detail. Due to the vast quantity of collected data and tremendous numerical calculations, a nesting technology is applied where the model goes, stepwise, from the global to the local scale. For each inserted model box, the lateral boundary values are derived from outer global or regional analysis data. Therefore, the WRF model also provides realistic input data for post-processing with conventional models concerning the accumulation of different types of atmospheric icing, including rime (in-cloud) icing, wet snow and freezing rain.
In principle, this only requires the data input on the standard weather interpolated in 3-D grids. However, since atmospheric icing is not a standard weather parameter, any type of direct and indirect measurement of icing will certainly improve the quality of the calculations.
Because atmospheric icing often occurs as a local phenomenon, and icing intensity varies greatly in space, especially in complex terrain, the modeling of icing requires a high horizontal resolution. To deal with this challenge, the model may apply grid spacing, often in the range of 0.4 km to 0.8 km (0.25 miles to 0.5 miles), which is considered an extremely high resolution for mesoscale models.
Application of Meteorological Model in Norway
The Norwegian Meteorological Institute has been extremely involved in the application of meteorological advancements in atmospheric icing. The first application of this approach was applied to a proposed route for a new 420-kV overhead transmission line in the western part of Norway, where a section of the line would be exposed to the North Sea at an altitude of 1,100 m (3,609 ft) and the risk of icing was expected to be very severe. The model setup included calibration with field measurements from the test site, located roughly 150 km (93 miles) south-southeast of the line route, in addition to some local measurements. The calculated ice loads on a theoretical vertical cylinder of 30 mm (1.2 inches) in diameter resulted in an ice load of 50 kg/m (33.6 lb/ft) on the reference rod during the test period.
This model was tested further within a European collaboration program on atmospheric icing on structures. It showed adequate performance for several European measuring sites, especially for rime icing in the Swiss Alps, but also for wet snow events on overhead line networks in the U.K.
Following these successful tests, the model was then applied on several overhead line projects in Greenland, Newfoundland (Canada), Chile and the U.K. Before setting up the model for local icing studies, it is first necessary to scan long-term series of standard meteorological data to identify potentially severe icing incidents. After studying a number of such events, it is then possible to evaluate icing severities along different line routes and also to estimate maximum ice loads that may have occurred over the time span covered by the available data.
Therefore, the final design loads may be selected with a much higher reliability than if the studies had not been done. The credibility of these load assessments will be further enhanced if such model studies can be combined with field measurements of ice loadings, for instance with ice racks and measurements of cloud water parameters.
Enhancing the WRF Model
The output from the WRF model can be embedded into Google Earth files. This creates a useful tool for visualization and makes it possible to move in and out of the landscape, and to see the local terrain in combination with the 3-D ice load outputs from any viewpoint of interest in each case. The value of this enhanced application was used in the Long Range Mountains in Newfoundland. It showed rime ice is dependent on the topography of the route, where valleys may enhance the lifting of moist air masses on the windward side and, hence, increase the risk of icing. Downdraft winds (subsidence) on the leeward side will dry out the clouds quickly and reduce the icing risk.
By studying the terrain, it is easy to identify at which levels the risk of rime icing may start for different line routes and different wind directions. Also, it is possible to optimize line routes to avoid the most severe icing areas and to evaluate the levels of icing over mountain plateaus and in mountain passes. Again, field measurements can improve these valuations.
This technology can be applied on case studies of historic events. In December 1990, there was a major storm in England with significant amounts of wet snow in the southern part of the Pennines. Approximately 250,000 customers lost power from the failure of about 700 high-voltage overhead line circuits and many low-voltage networks in the area of one distribution network operator (DNO). Studies indicated the corresponding radial ice thickness (RIT) of the accumulated wet snow was some 30 mm. According to reports from the U.K. DNO, the results compared very well with the DNO's observations and experiences from this event.
Similarly, another example from the U.K. centered on a field site for conductor testing and ice measurements located on the Deadwater Fell mountaintop, 580 m (1,903 ft) above sea level, near the England-Scotland border. The Deadwater Fell site is owned and operated by EA Technology. A significant rime icing event was recorded at this site in January 2010. EA Technology decided to test the site with the WRF model and extended the test to cover the British Isles. The model confirmed rime icing occurred over all high areas across the British Isles, with the highest loads on mountains recorded in Scotland, Northern Pennines, Wales and Ireland.
Today's Models Versus Previous Methods
There is no limit to the situations in which today's models for assessing rime icing in mountainous terrain can be applied. The current models confirm that the previous methods of assessing rime icing — extrapolating data from a regular weather station in the lowlands — are now regarded as inadequate. This is especially important because the temperature and cloud layers, in single cases, may vary differently in height than they do in average terms, often leading to unsatisfactory conclusions.
For wet snow, and possibly for freezing rain, the combination with a 3-D atmospheric weather forecasting model may expand the validity area of icing calculations based on single weather-station data. With more test cases and experience, it will be more useful and interesting to apply such fine-scale models of the atmosphere, especially for areas where no other data can be established.
The author is indebted to many people and institutions for their contributions to this article, including Bjørn Egil Kringlebotn Nygaard, a Ph.D. student who has made tremendous achievements to develop the WRF model for icing applications. The author also thanks EA Technology and Nalcor Energy for permission to refer to their projects, especially EA Technology for kind permission to publish the figure for the British Isles.
Svein M. Fikke (firstname.lastname@example.org) has a master's degree in meteorology and more than 35 years experience in extreme weather impacts on overhead power lines in exposed mountain climates. He was responsible for load assessments on the electricity networks in Norway until retirement from Statnett SF, the Norwegian Power Grid Co. Fikke now works as an independent consultant on international projects and also is active in international organizations, including CIGRÉ, IEC and ISO. He has published several articles on atmospheric icing on power networks.
EA Technology www.eatechnology.com
Nalcor Energy www.nalcorenergy.com
Norwegian Meteorological Institute www.met.no