above calculation method and software tool that will provide power generating wind farm, is working with Thomas Rubanovich Department of Electrical and Control Engineering Gdansk University of Technology. Its application can protect the owners of wind farms from paying liquidated damages resulting from erroneous assumptions. Accurate forecasts help control electricity network operators and optimal use of resources.
wind farm power depends largely on wind speed. Forecast errors combine the power estimation error and the error of forecast. Prediction of wind speed is very important, and achieving the objective of forecast error can improve the power steering rozpływami network.
"By partnering with investors will be able to refine the method is and build an application that in the future can be made available to every potential owner of a small and large wind farm - says in an interview with Thomas Rubanovich PAP.
As he explains, applications currently available are expensive and owners of small wind farms can not afford to buy them. Furthermore, they are inaccurate and errors in the forecasts reach 15-20 percent. Low accuracy of the forecasts by the company may have adverse economic effects resulting from the additional fees.
researcher intends to increase the accuracy of 24-hour forecasts using artificial neural networks. Seeking a response to the question of whether it is possible to forecast with an error 10 percent. based on a standard weather forecast.
Known tools to assist forecasting are based on a simple analytical model. It takes into account only the characteristics of the turbines, wind speed, air pressure and temperature, which does not provide the required accuracy. Slightly more factors include recursive models.
"The task becomes particularly difficult when we want to base the estimation of power at the weather forecast for a certain area and not a specific location" - states Rubanovich.
his view, the model predictions should take into account not only the profile of wind, wind direction, shadowing, but also the dynamics of changes in wind direction. Such a complex model requires a dynamic neural network structure accordingly.
scientist, prof. Associate. Dr PG. Eng. Elizabeth Bogaleckiej, conducts research related to the assessment of the reliability of production data and meteorological forecast model structure selection, evaluation of known analytical models, the term structure of the neural model, simulation and experiment.
Its purpose is to implement the method of calculation. The developed tool will be useful for those responsible for balancing the power grid and wind farm owners.
"Using tools built there is no limitation of regional concerns of each wind farm, the other is just a sample of learning, that is, the final parameters of the Web" - Thomas provides Rubanovich.
notes that his job is scientifically valid, as evidenced by numerous international publications and the fact that this topic is often discussed by researchers and investors at industry conferences.
0 comments:
Post a Comment