Unsupervised learning for residential energy consumption analytics
Abstract/Contents
- Abstract
- Developing strategies to encourage reduction in households energy consumption requires utilities to categorize, predict and modify consumers electricity usage. Unfortunately, a typical consumer exhibits wide variation in daily 24-hour electricity usage patterns. Traditional clustering methods have resulted in many hundreds of clusters, with a given consumer often associated with several clusters, making it difficult to classify consumers into stable representative groups and to predict individual electricity usage patterns. This dissertation presents two methods that better cluster consumer electricity usage pattern. The first method uses Dynamic time warping (DTW), which seeks an optimal alignment between electricity usage patterns. The second method assumes that electricity usage composes of a sequence of blocks generated from electrical devices. The clustering uses a novel block factorization model that embeds consumer usage pattern into a low dimensional space that flexibly captures time shifts of usage. Compared to commonly used clustering algorithm, both methods result in a more distinct set of clusters, and on average, a given consumer associates with a fewer clusters. The ideas and results from clustering is then used for individual electricity usage prediction. Prediction is done at two levels. The first level is day-to-day prediction where the shape of next day electricity usage pattern is predicted by cluster representatives from DTW. The second level is finer grain prediction based on block idea where load curve is predicted through block sequences. Predictions at both levels result in lower prediction error compared to some popular load forecasting techniques.
Description
| Type of resource | text |
|---|---|
| Form | electronic resource; remote; computer; online resource |
| Extent | 1 online resource. |
| Place | California |
| Place | [Stanford, California] |
| Publisher | [Stanford University] |
| Copyright date | 2018; ©2018 |
| Publication date | 2018; 2018 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Author | Teeraratkul, Thanchanok |
|---|---|
| Degree supervisor | Lall, Sanjay |
| Thesis advisor | Lall, Sanjay |
| Thesis advisor | O'Neill, Daniel C. (Daniel Craig) |
| Thesis advisor | Van Roy, Benjamin |
| Degree committee member | O'Neill, Daniel C. (Daniel Craig) |
| Degree committee member | Van Roy, Benjamin |
| Associated with | Stanford University, Department of Electrical Engineering. |
Subjects
| Genre | Theses |
|---|---|
| Genre | Text |
Bibliographic information
| Statement of responsibility | Thanchanok Teeraratkul. |
|---|---|
| Note | Submitted to the Department of Electrical Engineering. |
| Thesis | Thesis Ph.D. Stanford University 2018. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2018 by Thanchanok Teeraratkul
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...