dc.contributor.author |
Chopra, Abhinav Rakesh |
|
dc.contributor.author |
Nair, Nirmal Kumar |
|
dc.date.accessioned |
2021-08-04T23:58:14Z |
|
dc.date.available |
2021-08-04T23:58:14Z |
|
dc.date.issued |
2021-4-2 |
|
dc.identifier.isbn |
9781728186481 |
|
dc.identifier.uri |
https://hdl.handle.net/2292/55851 |
|
dc.description.abstract |
A wavelet decomposition based artificial neural network trained with an extreme learning machine is proposed for long-term load forecasting based on continuous learning. The proposed approach has been undertaken in pursuit of creating a software tool that can predict the impact of increased renewable distributed sources on load profiles. Several concerns and challenges arise as a result of this increase, as large transients are formed in these profiles causing a potential over-generation problem. As the number of smart meters and advanced metering infrastructure also increases, the demand for big data analytics systems becomes increasingly urgent. Pairing an intelligent hybrid neural network solution with an online cloud software makes the management of big data, in the form of smart meter load consumption data, more efficient and effective. The process of the proposed solution is that it first decomposes this data into components that represent more detailed periodic information within the time series. These components are then input into several extreme learning machine algorithms and the outputs are then recombined to make a load forecast and prediction of future load consumption data. The learning behaviour is vital and helps tuning with each new iteration increasing accuracy. A comparison of this paper's approach is benchmarked against the popular Levenberg-Marquardt training algorithm. The results obtained indicate the superiority of the wavelet-extreme learning machine hybrid compared to industry standard statistical methods and other hybrid neural network techniques. |
|
dc.publisher |
IEEE |
|
dc.relation.ispartof |
2021 IEEE Power and Energy Conference at Illinois (PECI) |
|
dc.relation.ispartofseries |
2021 IEEE Power and Energy Conference at Illinois (PECI) |
|
dc.rights |
Items in ResearchSpace are protected by copyright, with all rights reserved, unless otherwise indicated. Previously published items are made available in accordance with the copyright policy of the publisher. |
|
dc.rights.uri |
https://researchspace.auckland.ac.nz/docs/uoa-docs/rights.htm |
|
dc.title |
Wavelet-Extreme Learning i-Machine for New Zealand Smart Meter Data |
|
dc.type |
Conference Item |
|
dc.identifier.doi |
10.1109/peci51586.2021.9435262 |
|
pubs.begin-page |
1 |
|
pubs.volume |
00 |
|
dc.date.updated |
2021-07-05T10:13:51Z |
|
dc.rights.holder |
Copyright: The author |
en |
pubs.end-page |
8 |
|
pubs.finish-date |
2021-4-2 |
|
pubs.publication-status |
Published |
|
pubs.start-date |
2021-4-1 |
|
dc.rights.accessrights |
http://purl.org/eprint/accessRights/RestrictedAccess |
en |
pubs.elements-id |
857070 |
|