Machine Learning and Blockchain for Smart Factory Operations with Peer-to-Peer Transactive Energy Trading
| dc.contributor.advisor | Jin, Tongdan | |
| dc.contributor.author | Budde, Kathryn | |
| dc.contributor.committeeMember | Londa, Michelle | |
| dc.contributor.committeeMember | Zahed, Karim | |
| dc.date.accessioned | 2023-12-22T15:29:22Z | |
| dc.date.available | 2023-12-22T15:29:22Z | |
| dc.date.issued | 2023-12 | |
| dc.description.abstract | As renewable energy has become more prominent in the commercial and private sector, the use of third-party utility companies will likely become obsolete for those who can produce energy on site. With the accessibility of wind turbines, solar photovoltaic panels, and electric vehicles, peer-to-peer transactive energy will give factories and enterprise systems the ability to buy and sell onsite or microgrid energy at their own price. However, power intermittency and high upfront investment remain the main obstacle against large use of onsite renewable generation. This thesis aims to accomplish following research tasks. First, for energy producers in a smart factory system, machine learning algorithms are implemented to forecast the uncertain climate conditions that affect these renewable energy generation, hence balancing the power output and the demand for a day-ahead prediction. Second, to forecast the wind speed, recurrent neural networks and long short-term memory models are shown to achieve promising results for the regression analysis. Third, for the classification of the weather condition, K-nearest neighbor, decision tree, and random forest models provide accurate predictions as well. These models are tested intensively in Austin, TX and Boston, MA with seven years from 2010 to 2016. Given these forecasts, distributed ledger technologies like blockchain can be further adopted to assist in operating a decentralized energy market. Manifested as smart contracts, blockchain facilitates the decentralized two-way energy transactions between individual industrial facilities or residential homes where the generation of excess renewable or microgrid energy will now be profitable. | |
| dc.description.department | Engineering | |
| dc.format | Text | |
| dc.format.extent | 121 pages | |
| dc.format.medium | 1 file (.pdf) | |
| dc.identifier.citation | Budde, K. (2023). Machine learning and blockchain for smart factory operations with peer-to-peer transactive energy trading (Unpublished thesis). Texas State University, San Marcos, Texas. | |
| dc.identifier.uri | https://hdl.handle.net/10877/17820 | |
| dc.language.iso | en | |
| dc.subject | machine learning | |
| dc.subject | blockchain | |
| dc.subject | photovoltaics | |
| dc.subject | wind turbines | |
| dc.subject | long short term memory | |
| dc.subject | recurrent neural network | |
| dc.subject | regression | |
| dc.subject | classification | |
| dc.subject | transactive | |
| dc.subject | energy | |
| dc.subject | distributed generation | |
| dc.title | Machine Learning and Blockchain for Smart Factory Operations with Peer-to-Peer Transactive Energy Trading | |
| dc.type | Thesis | |
| thesis.degree.department | Ingram School of Engineering | |
| thesis.degree.discipline | Engineering | |
| thesis.degree.grantor | Texas State University | |
| thesis.degree.level | Masters | |
| thesis.degree.name | Master of Science |