Estimation of Resource Allocation Based on Disturbance Prediction Data with Use of Statistics, Machine Learning and Data Analysis
(2017) In Master's Theses in Mathematical Sciences FMA820 20171Mathematics (Faculty of Engineering)
- Abstract
- The electrical power grid is one of modern society’s most important infrastructures and both power distributors and the Swedish government are investing large amount of resources to ensure continuous delivery of power. By predicting future outages with an automatic prediction system, the distributors could prevent long restoration times and economic loss. In this work, the authors evaluate the possibility of predicting outages based on statistical and machine learning methods and the relative importance of the different factors. The study uses open weather data and data on the power grid gathered from Swedish Meteorological Institute and E.ON, respectively.
The result shows that while maintaining the same false positive rate as... (More) - The electrical power grid is one of modern society’s most important infrastructures and both power distributors and the Swedish government are investing large amount of resources to ensure continuous delivery of power. By predicting future outages with an automatic prediction system, the distributors could prevent long restoration times and economic loss. In this work, the authors evaluate the possibility of predicting outages based on statistical and machine learning methods and the relative importance of the different factors. The study uses open weather data and data on the power grid gathered from Swedish Meteorological Institute and E.ON, respectively.
The result shows that while maintaining the same false positive rate as currently used manual prediction methods, the automatic prediction is able to increase the true positive rate from 20% to 33%. The authors conclude that wind gust is the most important factor in predicting weather related outages and that the models are better to predict outages during stronger winds. However, further data analysis is warranted before the automatic prediction can be implemented in a real world context. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/8907090
- author
- Domeij, Rebecka LU and Luong, Richard LU
- supervisor
- organization
- course
- FMA820 20171
- year
- 2017
- type
- H2 - Master's Degree (Two Years)
- subject
- keywords
- outages, prediction, negative binomial regression, logistic regression, decision tree
- publication/series
- Master's Theses in Mathematical Sciences
- report number
- LUTFMA-3316-2017
- ISSN
- 1404-6342
- other publication id
- 2017:E19
- language
- English
- id
- 8907090
- date added to LUP
- 2017-05-29 16:22:19
- date last changed
- 2017-05-29 16:22:19
@misc{8907090, abstract = {{The electrical power grid is one of modern society’s most important infrastructures and both power distributors and the Swedish government are investing large amount of resources to ensure continuous delivery of power. By predicting future outages with an automatic prediction system, the distributors could prevent long restoration times and economic loss. In this work, the authors evaluate the possibility of predicting outages based on statistical and machine learning methods and the relative importance of the different factors. The study uses open weather data and data on the power grid gathered from Swedish Meteorological Institute and E.ON, respectively. The result shows that while maintaining the same false positive rate as currently used manual prediction methods, the automatic prediction is able to increase the true positive rate from 20% to 33%. The authors conclude that wind gust is the most important factor in predicting weather related outages and that the models are better to predict outages during stronger winds. However, further data analysis is warranted before the automatic prediction can be implemented in a real world context.}}, author = {{Domeij, Rebecka and Luong, Richard}}, issn = {{1404-6342}}, language = {{eng}}, note = {{Student Paper}}, series = {{Master's Theses in Mathematical Sciences}}, title = {{Estimation of Resource Allocation Based on Disturbance Prediction Data with Use of Statistics, Machine Learning and Data Analysis}}, year = {{2017}}, }