Weather
Weather tech wind power generation in cold climates may be greatly affected by icing, so it is of considerable importance to know how icing conditions will be at sites chosen for possible wind farms. Past research has found that if the numerical models employed for estimating icing conditions over regions of complex terrain are to be used, they must be run at very high resolution, as low as one kilometer.
Currently, running such high-resolution models over a region the size of Sweden is not possible for extended periods, say 30 years. To that end some procedure for selecting representative periods will need to be designed. Toward this, various classifying algorithms will be used for the re-analysis of data. ERA-interim re-analysis data of ECMWF include a global analysis atmospheric and surface weather parameters from 1989 to the present with a resolution of 6 hours in time.
The horizontal scale is approximately 80 km. Up until now, the Lamb classification technique has been tried out on this dataset to identify representative months. Mean sea level pressure analysis is employed in classifying weather conditions. ERA-interim data has also been utilized to compute ice loads, applying the Makkonen formula as outlined. Initial comparisons with ice load observations at a few locations in northern Sweden indicate that icing event timing is well captured but not the load levels.
Three various high-resolution mesoscale models have been executed in a number of configurations throughout the icing season 2009/2010. The models are COAMPS Coupled Ocean/Atmosphere Mesoscale Prediction System, weather parameters, weather tech, weather tech floor mats, weather tech com, weather tech coupons AROME Applications of Research to Operations at Mesoscale and the WRF Research and Forecasting model. Model run output has been utilized to compute the ice load using the Makkonen formula.
Observations
A minor network of weather tech icing measurement points was set up in the northern region of Sweden during the winter period 2009/2010. The principal ice load-measuring device on the stations is the SAAB Security Ice Monitor. This device accumulates the ice on a 0.5-meter vertically mounted rotating cylinder, and the accumulated ice is weighed continuously. An optical sensor by Holooptics has also been employed.
This sensor currently only senses ice yes/no, weather tech floor mats and we do not receive any accumulations. In addition to icing, usual meteorological parameters such as wind speed, wind direction, pressure, temperature, and humidity are also sensed. The sensors are placed either on a wind power plant or a telecommunication tower. We had three locations online last winter, and thus far this winter, two more stations have been set up, and another six are being sent out. To some of the locations, observations of visibility and cloud base are also included in the above parameters. The Ice Monitor bottom and Holooptics top from the Sveg measurement location.
Ice accretion model
The result from weather tech mesoscale model simulations is applied in the calculation of ice loads using a cylindrical model of ice accretion, commonly known as the Makkonen model. Temperature, wind speed, supercooled liquid water content, and median volume droplet size are the inputs to this model. The latter cannot be obtained from the current models, so a prescribed constant has to be used.
Time series of the atmospheric variables can be cut out from our model run at a one-hour time resolution and be inserted into the accretion model, and as a result, we have an ice load time series to compare with observations. Even if the models are of high resolution, tech weather floor mats, weather guard trucks, winter tech mats, weather tech near me, weather tech car floor we are still missing some topographic resolution, so mountain tops in the models are, in most cases, below the actual tops.
To make up for this height difference, the model data temperature and specific humidity are raised adiabatically to the observation height. Condensation is possible during this process if it happens, the condensed water is added to the model liquid water before the accretion calculation.
ERA-interim
The ECMWF ERA-interim weather tech re-analysis data provide the entire globe from 1989 until now with a 6-hour time resolution. The horizontal resolution is approximately 80 kilometers, and there are 60 vertical levels in the air. All parameters that can be measured at the surface, as well as in the free atmosphere, have been included in the database. The project will utilize this data in two ways: first, as representative icing periods, which will be run with high resolution, and second, as initial conditions and lateral boundaries to the high-resolution runs.
Some initial tests have been conducted to determine the representative periods using an automatic version of the Lamb weather classification. In this categorization, a low-resolution analysis of mean sea level pressure for each day is employed to categorize the conditions.
Examining the relationship between the geostrophic wind and the vorticity of the pressure field produces as an outcome a weather class that may be low pressure, high pressure, or one of eight prevailing wind directions. Hybrid classes involve low/high pressure combined with wind direction.
The classification has been performed for three regions of Sweden northern, middle, and southern. The mean values of the distribution of classes for each region and each month of winter have been computed, as well as which month was closest to this mean distribution. These months would then be potential candidates for high-resolution simulations. It is not known, though, whether this is the best method to select further research will be conducted.
CONCLUSIONS
An effort has been made to weather tech generate a high-resolution icing climatology for Sweden. This is an ongoing work, and no end product has been shown to date. The facilities to generate this climatology have been validated within this ongoing project, which will continue for another two years. State-of-the-art high-resolution mesoscale weather models are the most significant tool in our opinion.
Experience to date, comparing modeled ice load to observations, indicates that we must use a horizontal resolution of approximately 1 kilometer for the runs of the models. Because such a model run is highly computer-intensive for large domains, we must somehow select representative periods for those runs. The representative periods will be determined based on global reanalysis data.