Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
Optimizing Battery Storage for Minimizing Forecast Errors in Wind Power Systems
van der Merwe, S. W. 2025. Optimizing Battery Storage for Minimizing Forecast Errors in Wind Power Systems. Unpublished masters thesis. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/38dabddf-409e-4f19-8cb1-cf8dbff39ab5
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
Stellenbosch : Stellenbosch University
2025
|
| Subjects: | |
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867613820510273536 |
|---|---|
| access_status_str | Open Access |
| author | Van der Merwe, Schalk Willem |
| author2 | Rix, Arnold J. |
| author_browse | Rix, Arnold J. Van der Merwe, Schalk Willem |
| author_facet | Rix, Arnold J. Van der Merwe, Schalk Willem |
| author_sort | Van der Merwe, Schalk Willem |
| collection | Thesis |
| dc_rights_str_mv | Stellenbosch University |
| description | van der Merwe, S. W. 2025. Optimizing Battery Storage for Minimizing Forecast Errors in Wind Power Systems. Unpublished masters thesis. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/38dabddf-409e-4f19-8cb1-cf8dbff39ab5 |
| format | Thesis |
| id | oai:scholar.sun.ac.za:10019.1/132364 |
| institution | Stellenbosch University (South Africa) |
| language | English |
| last_indexed | 2026-06-10T12:42:12.448Z |
| license_str | Other — see source repository |
| provenance_str_mv | Harvested via OAI-PMH from SUNScholar — Stellenbosch University Repository |
| publishDate | 2025 |
| publishDateRange | 2025 |
| publishDateSort | 2025 |
| publisher | Stellenbosch : Stellenbosch University |
| publisherStr | Stellenbosch : Stellenbosch University |
| record_format | dspace |
| source_str | SUNScholar — Stellenbosch University Repository |
| spelling | oai:scholar.sun.ac.za:10019.1/132364 Optimizing Battery Storage for Minimizing Forecast Errors in Wind Power Systems Van der Merwe, Schalk Willem Rix, Arnold J. Du Plessis, Armand A. Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering. Wind power -- Forecasting -- South Africa Energy storage -- South Africa Photovoltaic power systems -- South Africa Mathematical optimization UCTD van der Merwe, S. W. 2025. Optimizing Battery Storage for Minimizing Forecast Errors in Wind Power Systems. Unpublished masters thesis. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/38dabddf-409e-4f19-8cb1-cf8dbff39ab5 Thesis (MEng)--Stellenbosch University, 2025. ENGLISH ABSTRACT: With the increase in global energy demand and the implementation of load shedding in South Africa, the need for new energy sources is more pressing than ever. Renewable energy is central to discussions about the future of energy production. Wind energy has gained popularity among renewable sources compared to solar photovoltaic (PV) and bioenergy due to its high efficiency, abundance, low environmental impact, and ability to generate power independently of sunlight. However, integrating wind energy into the existing electrical grid presents additional challenges. The inherent stochastic nature of wind necessitates accurate wind power forecasting models to ensure reliable and widespread integration into the grid. This thesis explores how to forecast wind power effectively and with appropriate methods. It details a systematic approach for developing feed-forward neural network models tailored to wind power prediction for a new location. Using extensive historical meteorological data and advanced feature engineering techniques, a novel input feature was developed for forecasting. A grid search algorithm was employed to identify the optimal model architecture, with linear regression models used as benchmarks to evaluate the final models. The selection process was guided by accuracy and computational efficiency. The research also includes a quantitative analysis comparing single-step and multi-step models, highlighting a preference for the latter due to their more comprehensive forecasting capabilities. Additionally, the thesis investigates the computational effects and accuracy impacts of different training step sizes. With this research, a software battery model and simulation to accurately size a battery energy storage system for minimizing errors in wind power forecasting models. The simulation successfully sized battery systems to achieve 70%, 80%, and 90% error reduction, demonstrating the system’s effectiveness in mitigating forecasting errors at these levels. Additionally, the simulation revealed that the battery energy storage system is more effective at mitigating numerous smaller errors compared to fewer, larger-magnitude errors. A case study was conducted to investigate the effect of optimizing the wind turbine and battery energy storage system for the wind characteristics of the tested location. The optimisation case study detailed the methodology and results obtained, highlighting the trade-off between smaller wind turbines and battery sizes. AFRIKAANSE OPSOMMING: Met die toename in wˆereldwye energie aanvraag, asook die implimentering van beurtkrag in Suid-Afrika, is die behoefte vir nuwe energiebronne meer dringend as ooit. "Hernubare energie"staan sentraal in besprekings oor die toekoms van energieproduksie. Windenergie het gewildheid verwerf onder hernubare bronne in vergelyking met sonfotovoltaïese en bio-energie, weens sy doeltreffendheid, oorvloed, lae negatiewe omgewingsimpak en vermoë om krag onafhanklik van sonlig te genereer. Die integrasie van windenergie in die bestaande elektrisiteitsnetwerk bied egter bykomende uitdagings. Die inherente stogastiese aard van wind vereis akkurate windkragvoorspellingsmodelle om betroubare en wydverspreide integrasie in die netwerk te verseker. Hierdie navorsing ondersoek hoe om windkrag effektief en met toepaslike metodes te voorspel. Dit beskryf ’n sistematiese benadering vir die ontwikkeling van Vorentoevoer neurale netwerkmodelle wat spesifiek aangepas is vir windkragvoorspelling by ’n unieke ligging. Deur uitgebreide historiese meteorologiese data en gevorderde kenmerkingenieurswese-tegnieke te gebruik, is ’n nuwe invoerkenmerk ontwikkel vir voorspellingsdoeleindes. ’n Rooster-soekalgoritme is gebruik om die optimale modelargitektuur te identifiseer, met lineêre regressiemodelle wat as maatstawwe gebruik is om die finale modelle te evalueer. Die keuseproses is gelei deur akkuraatheid en rekenaar-effektiwiteit. Die navorsing sluit ook ’n kwantitatiewe ontleding in wat enkel-stap en multi-stap modelle vergelyk, en ’n voorkeur vir laasgenoemde uitlig weens hul meer omvattende voorspellingsvermoëns. Daarbenewens ondersoek die tesis die rekenaareffekte en akkuraatheidsimpakte van verskillende opleidingsstapgroottes. Hierdie tesis het ook ’n sagteware-batterymodel en simulasie ontwikkel om die grootte van ’n battery-energiebergingstelsel akkuraat te bepaal met die doel om foute in windkragvoorspellingsmodelle te minimeer. Die simulasie het suksesvol batterystelsels in grootte aangepas om 70%, 80% en 90% foutvermindering te bereik, wat op sig self die stelsel se doeltreffendheid gedemonstreer het in die vermindering van voorspellingsfoute op hierdie vlakke. Verder het die simulasie gewys dat die battery-energiebergingstelsel meer effektief is in die vermindering van talle kleiner foute in vergelyking met minder, groter foute. ’n Gevallestudie is uitgevoer om die effek van die optimering van die windturbine en battery-energiebergingstelsel vir die windkenmerke van die getoetste ligging te ondersoek. Die optimiseringsgevalstudie het die metodiek en verkrygde resultate uiteengesit, en die kompromie tussen kleiner windturbines en battery groottes uitgelig. Masters 2025-06-05T06:53:11Z 2025-06-05T06:53:11Z 2025-03 Thesis https://scholar.sun.ac.za/handle/10019.1/132364 en Stellenbosch University xviii, 168 pages : ill.ustration application/pdf Stellenbosch : Stellenbosch University |
| spellingShingle | Wind power -- Forecasting -- South Africa Energy storage -- South Africa Photovoltaic power systems -- South Africa Mathematical optimization UCTD Van der Merwe, Schalk Willem Optimizing Battery Storage for Minimizing Forecast Errors in Wind Power Systems |
| title | Optimizing Battery Storage for Minimizing Forecast Errors in Wind Power Systems |
| title_full | Optimizing Battery Storage for Minimizing Forecast Errors in Wind Power Systems |
| title_fullStr | Optimizing Battery Storage for Minimizing Forecast Errors in Wind Power Systems |
| title_full_unstemmed | Optimizing Battery Storage for Minimizing Forecast Errors in Wind Power Systems |
| title_short | Optimizing Battery Storage for Minimizing Forecast Errors in Wind Power Systems |
| title_sort | optimizing battery storage for minimizing forecast errors in wind power systems |
| topic | Wind power -- Forecasting -- South Africa Energy storage -- South Africa Photovoltaic power systems -- South Africa Mathematical optimization UCTD |
| url | https://scholar.sun.ac.za/handle/10019.1/132364 |
| work_keys_str_mv | AT vandermerweschalkwillem optimizingbatterystorageforminimizingforecasterrorsinwindpowersystems |