Preprint Article Version 1 This version is not peer-reviewed

# Forecast Horizon and Solar Variability Influences on the Performances of Multiscale Hybrid Forecast Model (MHFM)

Version 1 : Received: 4 September 2018 / Approved: 5 September 2018 / Online: 5 September 2018 (06:22:35 CEST)

How to cite: Monjoly, S.; Calif, R.; André, M.; Soubdhan, T. Forecast Horizon and Solar Variability Influences on the Performances of Multiscale Hybrid Forecast Model (MHFM). Preprints 2018, 2018090089 (doi: 10.20944/preprints201809.0089.v1). Monjoly, S.; Calif, R.; André, M.; Soubdhan, T. Forecast Horizon and Solar Variability Influences on the Performances of Multiscale Hybrid Forecast Model (MHFM). Preprints 2018, 2018090089 (doi: 10.20944/preprints201809.0089.v1).

## Abstract

In this paper, the forecast horizon and solar variability influences on MHFM model based on multiscale decomposition, AR and NN models, are studied. This article follows the works published in [1] showing the performance of the MHFM using 3 multiscale decomposition methods and a forecast horizon equal to 1 hour. Several forecast horizon strategies and his influence on the MHFM performances are investigated. We show that the best strategy for a rRMSE variying from $4.43\%$ to $10.24\%$ is obtained for forecast horizons from $5$ minutes to $6$ hours. In a second part, the solar variability influence on the MHFM is studied. A classification based on a shows that the best performance of MHFM is obtained for clear sky days with a rRMSE of $2.91\%$ and worst for cloudy sky days with a rRMSE of $6.73\%$.

## Subject Areas

hybrid forecast model; forecast horizon; daily global solar radiation clustering; fuzzy c-means; variability characterization