Predictive and Prescriptive Analytics for Managing the Impact of Hazards on Power Systems
dc.contributor.author | Kabir, Elnaz | |
dc.date.accessioned | 2021-06-08T23:12:45Z | |
dc.date.available | 2021-06-08T23:12:45Z | |
dc.date.issued | 2021 | |
dc.date.submitted | 2021 | |
dc.identifier.uri | https://hdl.handle.net/2027.42/168024 | |
dc.description.abstract | Natural hazards and extreme weather events have the potential to cause significant disruptions to the electric power grid. The resulting damages are, in some cases, very expensive and time-consuming to repair and they lead to substantial burdens on both utilities and customers. The frequency of such events has also been increasing over the last 30 years and several studies show that both the number and intensity of severe weather events will increase due to global warming and climate change. An important part of managing weather-induced power outages is being properly prepared for them, and this is tied in with broader goals of enhancing power system resilience. Inspired by these challenges, this thesis focuses on developing data-driven frameworks under uncertainty for predictive and prescriptive analytics in order to address the resiliency challenges of power systems. In particular, the primary aims of this dissertation are to: 1. Develop a series of predictive models that can accurately estimate the probability distribution of power outages in advance of a storm. 2. Develop a crew coordination planning model to allocate repair crews to areas affected by hazards in response to the uncertain predicted outages. The first chapter introduces storm outage management and explains the main objectives of this thesis in detail. In the second chapter, I develop a novel two-stage predictive modeling framework to overcome the zero-inflation issue that is seen in most outage related data. The proposed model accurately estimates customer interruptions in terms of probability distributions to better address inherent stochasticity in predictions. In the next chapter, I develop a new adaptive statistical learning approach based on Bayesian model averaging to formulate model uncertainty and develop a model that is able to adapt to changing conditions and data over time. The forth chapter uses Bayesian belief network to model the stochastic interconnection between various meteorological factors and physical damage to different power system assets. Finally, in chapter five, I develop a new multi-stage stochastic program model to allocate and relocate repair crews in impacted areas during an extreme weather event to restore power as quickly as possible with minimum costs. This research was conducted in collaboration with multiple power utility companies, and some of the models and algorithms developed in this thesis are already implemented in those companies and utilized by their employees. Based on actual data from these companies, I provide evidence that significant improvements have been achieved by my models. | |
dc.language.iso | en_US | |
dc.subject | Predictive & Prescriptive Analytics for Managing the Impact of Hazards on Power Systems | |
dc.title | Predictive and Prescriptive Analytics for Managing the Impact of Hazards on Power Systems | |
dc.type | Thesis | |
dc.description.thesisdegreename | PhD | en_US |
dc.description.thesisdegreediscipline | Industrial & Operations Engineering | |
dc.description.thesisdegreegrantor | University of Michigan, Horace H. Rackham School of Graduate Studies | |
dc.contributor.committeemember | Guikema, Seth David | |
dc.contributor.committeemember | Gagnon Bartsch, Johann | |
dc.contributor.committeemember | Denton, Brian | |
dc.contributor.committeemember | Quiring, Steven | |
dc.subject.hlbsecondlevel | Industrial and Operations Engineering | |
dc.subject.hlbtoplevel | Engineering | |
dc.description.bitstreamurl | http://deepblue.lib.umich.edu/bitstream/2027.42/168024/1/ekabir_1.pdf | |
dc.identifier.doi | https://dx.doi.org/10.7302/1451 | |
dc.identifier.orcid | 0000-0002-5211-7433 | |
dc.identifier.name-orcid | Kabir, Elnaz; 0000-0002-5211-7433 | en_US |
dc.working.doi | 10.7302/1451 | en |
dc.owningcollname | Dissertations and Theses (Ph.D. and Master's) |
Files in this item
Remediation of Harmful Language
The University of Michigan Library aims to describe library materials in a way that respects the people and communities who create, use, and are represented in our collections. Report harmful or offensive language in catalog records, finding aids, or elsewhere in our collections anonymously through our metadata feedback form. More information at Remediation of Harmful Language.
Accessibility
If you are unable to use this file in its current format, please select the Contact Us link and we can modify it to make it more accessible to you.