The icing loss tab allows users to specify an icing loss model to be used in time-series energy capture calculations. This can either be loaded in from an external look-up table or it can be defined as a more physics-inspired model.
Look-up Table: this table takes the form of a tab-delimited text file with the following format.
The first line is ignored.
Speed is in m/s; temperature can be in degrees Celsius or Kelvin; relative humidity is in percent; Cumulative Precipitation is in mm and is calculated as a running total of the of the continuous precipitation and is only reset if three successive time-steps have zero precipitation.
The icing loss takes the form of an effective speed penalty which is contained in the last field and is the output corresponding to the four input fields.
The ability to import a lookup table means that the user can bring in any icing model whose inputs are velocity, temperature, humidity and precipitation or a subset of those. Once loaded, the lookup table can be explored using the drop-down lists in the Loaded Table box.
An alternative to loading in your own icing model as a lookup table is to use or modify the simple physical model that is shown in figure 111. The loss values in each grid are multiplied together along with the scale factor specified above the grids. The result is the effective speed penalty which is then combined with the power curve to come up with an energy loss.
Whichever icing loss model is used, it the resulting ESP (effective speed penalty) for any time-step in which icing is predicted to be detected (ESP>0) can be used simply as the basis for an energy loss but it can also be used to switch on blade heaters or to derate the capacity of the turbine.
Icing detected causes turbine derating – when checked, users can input a threshold for a requisite level of ESP (>=0) that then triggers the turbine to derate. An ESP of just above zero means that a small amount of icing has been detected which is enough to degrade the performance of the turbine but perhaps not enough to warrant derating.
Icing detected causes blade heating – when checked, users can set a threshold for the level of ESP which constitutes sufficient icing being detected so that the blade heating system is switched on. Blade heaters are modelled as either fully on or fully off. The energy consumed by them is not added to the icing loss but added to the consumption loss instead. Blade heating can cause the power production of the turbine to drop below zero for a given time-step. Blade heating causes the energy losses due to icing to be reduced by a user-defined percentage. At present, Openwind is not modelling full shutdowns caused by icing (which may be mitigated by blade heating systems).
NB: The icing loss model is very much a work in progress just now and is provided in the Enterprise version of Openwind to allow us to test but also to allow users to experiment and provide feedback. All aspects of this functionality are subject to change as we validate and refine this approach (one area which merits more attention is the de-icing rate). We hope to continually improve this functionality in upcoming versions.