Modern deep neural networks for Direct Normal Irradiance forecasting: A classification approach

The escalating energy demand and the adverse environmental impacts of fossil-fuel use necessitate a shift towards cleaner and renewable alternatives. Concentrated Solar Power (CSP) technology emerges as a promising solution, offering a carbon-free alternative for power generation. The efficiency and...

Full description

Saved in:
Bibliographic Details
Main Authors: Muhammad Saud Ul Hassan, Kashif Liaqat, Laura Schaefer, Alexander J. Zolan
Format: Article
Language:English
Published: Elsevier 2024-12-01
Series:e-Prime: Advances in Electrical Engineering, Electronics and Energy
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2772671124004327
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850256239665086464
author Muhammad Saud Ul Hassan
Kashif Liaqat
Laura Schaefer
Alexander J. Zolan
author_facet Muhammad Saud Ul Hassan
Kashif Liaqat
Laura Schaefer
Alexander J. Zolan
author_sort Muhammad Saud Ul Hassan
collection DOAJ
description The escalating energy demand and the adverse environmental impacts of fossil-fuel use necessitate a shift towards cleaner and renewable alternatives. Concentrated Solar Power (CSP) technology emerges as a promising solution, offering a carbon-free alternative for power generation. The efficiency and profitability of CSP depend on the Direct Normal Irradiance (DNI) component of solar radiation; hence, accurate DNI forecasting can help optimize CSP plants’ operations and performance. The unpredictable nature of weather phenomena, particularly cloud cover, introduces uncertainty into DNI projections. Existing DNI forecasting models use meteorological factors, which are both challenging to estimate numerically over short prediction windows and expensive to model through data at a sufficiently high spatial and temporal resolution. This research addresses the challenge by presenting a novel approach that formulates DNI prediction as a multi-class classification problem, departing from conventional regression-based methods. The primary objective of this classification framework is to identify optimal periods aligning with specific operational thresholds for CSP plants, contributing to enhanced dispatch optimization strategies. We model the DNI classification problem using four advanced deep neural networks – rectified linear unit (ReLU) networks, 1D residual networks (ResNets), bidirectional long short-term memory (BiLSTM) networks, and transformers – achieving accuracies up to 93.5% without requiring meteorological parameters.
format Article
id doaj-art-eb865a2aae584ee08314ea4e374dd244
institution OA Journals
issn 2772-6711
language English
publishDate 2024-12-01
publisher Elsevier
record_format Article
series e-Prime: Advances in Electrical Engineering, Electronics and Energy
spelling doaj-art-eb865a2aae584ee08314ea4e374dd2442025-08-20T01:56:41ZengElseviere-Prime: Advances in Electrical Engineering, Electronics and Energy2772-67112024-12-011010085310.1016/j.prime.2024.100853Modern deep neural networks for Direct Normal Irradiance forecasting: A classification approachMuhammad Saud Ul Hassan0Kashif Liaqat1Laura Schaefer2Alexander J. Zolan3Department of Mechanical Engineering, Rice University, Houston TX, 77005, United StatesDepartment of Mechanical Engineering, Rice University, Houston TX, 77005, United States; Corresponding author.Department of Mechanical Engineering, Rice University, Houston TX, 77005, United StatesThermal Energy Systems, National Renewable Energy Laboratory, Golden CO, 80401, United StatesThe escalating energy demand and the adverse environmental impacts of fossil-fuel use necessitate a shift towards cleaner and renewable alternatives. Concentrated Solar Power (CSP) technology emerges as a promising solution, offering a carbon-free alternative for power generation. The efficiency and profitability of CSP depend on the Direct Normal Irradiance (DNI) component of solar radiation; hence, accurate DNI forecasting can help optimize CSP plants’ operations and performance. The unpredictable nature of weather phenomena, particularly cloud cover, introduces uncertainty into DNI projections. Existing DNI forecasting models use meteorological factors, which are both challenging to estimate numerically over short prediction windows and expensive to model through data at a sufficiently high spatial and temporal resolution. This research addresses the challenge by presenting a novel approach that formulates DNI prediction as a multi-class classification problem, departing from conventional regression-based methods. The primary objective of this classification framework is to identify optimal periods aligning with specific operational thresholds for CSP plants, contributing to enhanced dispatch optimization strategies. We model the DNI classification problem using four advanced deep neural networks – rectified linear unit (ReLU) networks, 1D residual networks (ResNets), bidirectional long short-term memory (BiLSTM) networks, and transformers – achieving accuracies up to 93.5% without requiring meteorological parameters.http://www.sciencedirect.com/science/article/pii/S2772671124004327Direct Normal IrradianceConcentrated solar powerDeep neural networksRecurrent neural networksTransformers
spellingShingle Muhammad Saud Ul Hassan
Kashif Liaqat
Laura Schaefer
Alexander J. Zolan
Modern deep neural networks for Direct Normal Irradiance forecasting: A classification approach
e-Prime: Advances in Electrical Engineering, Electronics and Energy
Direct Normal Irradiance
Concentrated solar power
Deep neural networks
Recurrent neural networks
Transformers
title Modern deep neural networks for Direct Normal Irradiance forecasting: A classification approach
title_full Modern deep neural networks for Direct Normal Irradiance forecasting: A classification approach
title_fullStr Modern deep neural networks for Direct Normal Irradiance forecasting: A classification approach
title_full_unstemmed Modern deep neural networks for Direct Normal Irradiance forecasting: A classification approach
title_short Modern deep neural networks for Direct Normal Irradiance forecasting: A classification approach
title_sort modern deep neural networks for direct normal irradiance forecasting a classification approach
topic Direct Normal Irradiance
Concentrated solar power
Deep neural networks
Recurrent neural networks
Transformers
url http://www.sciencedirect.com/science/article/pii/S2772671124004327
work_keys_str_mv AT muhammadsaudulhassan moderndeepneuralnetworksfordirectnormalirradianceforecastingaclassificationapproach
AT kashifliaqat moderndeepneuralnetworksfordirectnormalirradianceforecastingaclassificationapproach
AT lauraschaefer moderndeepneuralnetworksfordirectnormalirradianceforecastingaclassificationapproach
AT alexanderjzolan moderndeepneuralnetworksfordirectnormalirradianceforecastingaclassificationapproach