Publication details
Analysis of Photovoltaic Energy Using Kernel Smoothing
| Authors | |
|---|---|
| Year of publication | 2015 |
| Type | Appeared in Conference without Proceedings |
| MU Faculty or unit | |
| Citation | |
| Description | The share of power generated by the solar plant is constantly increasing, despite the unreliable character of this type of energy. In order to ensure the optimal utilization of the generated power, its forecasting becomes more and more important. The crucial issue is to determine the relationship between weather characteristics, mainly solar radiation and temperature, and the photovoltaic energy. We propose to use a multivariate kernel regression for this task. Methods of kernel estimations represent one of the most effective smoothing techniques. These methods are simply to understand and they possess good statistical properties. The most important factor in the multivariate kernel regression is a bandwidth matrix. This matrix controls both the amount and the direction of the multivariate smoothing. Considerable attention is paid to constrained parametrization of the bandwidth matrix such as a diagonal matrix.We propose a method based on an optimally balanced relation between the integrated variance and the integrated squared bias. This method is also suitable for a gradient estimation, i.e., it is able to detect „gaps“ and „holes“ of the regression surface. In the present contribution the method is applied on data consisting of hourly values of energy generated by solar power plants in a specific area and of predictions of solar radiation and temperature provided by numerical prediction model ALADIN. |
| Related projects: |