Proposal of two new criteria to assess solar forecasting performances

International audience; A growing interest concerns the study of the temporal variability of solar irradiance which operates at different time and space scales. The exploitation and management of solar energy are directly impacted by this variability and may have negative impacts on the grid or on photovoltaic power plants that have means of storage or sell on the spot market. The objective is then to anticipate, at a given space scale, the temporal fluctuations of the incident solar irradiance which explain the variability on the electrical production. An increasing number of solar forecasting methods have been published in the literature to cope with this issue at different time horizons. To compare the performances of the methods, it is necessary to have a range of metrics along with a standardized procedure, allowing characterizing and comparing the performances of these forecasts. This poster illustrates some limitations of the widely used metric RMSE on a concrete example. In this work we propose a review of the literature related to metrics for solar forecasting, and we propose two new metrics that we deem relevant and innovative: a new ramp and time alignment metric.