Forecasting intra-hour variance of photovoltaic power using a new integrated model

Date

2021

Authors

Guermoui, M.
Bouchouicha, K.
Bailek, N.
Boland, J.W.

Editors

Advisors

Journal Title

Journal ISSN

Volume Title

Type:

Journal article

Citation

Energy Conversion and Management, 2021; 245(114569)

Statement of Responsibility

Conference Name

Abstract

Photovoltaic (PV) solar power, which is considered as the most competitive clean energy source, contributes to a significant percentage of electricity production in many developed countries. However, accurate PV power forecasting is necessary due to its high variation that can be caused by several factors. Hence, the intermittent nature of PV production represents a major challenge to integrate PV systems into the electric grid. The scope of this paper deals with this issue through developing a new integrated PV power forecasting model. The proposed model is based on the use of a new decomposition methodology, named Iterative Filtering for decomposing PV power into different intrinsic functions (IMFs), then Extreme Learning Machine (ELM) is used as essence predictor. To this end, the proposed IF-ELM model is evaluated on three solar PV power plants installed at three different sites, with different climatic conditions. Direct and recursive IF-ELM methodologies are examined for multi-step ahead forecasting in a very short time-scale (up to 60 min). Overall, the forecasting results show high precision performance for the studied forecasting horizons- in terms of different statistical metrics compared to stand-alone models. Also, the proposed IF method shows its high performance when compared to the recently developed decomposition method. complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) in improving the forecasting accuracy of a single model. Forecasting results with the IF-ELM model led to an error in nRMSE that is less than 10% and a Correlation Coefficient greater than 98% over all forecasting horizons.

School/Discipline

Dissertation Note

Provenance

Description

Access Status

Rights

Copyright 2021 Elsevier

License

Grant ID

Call number

Persistent link to this record