Matching Items (3)
Filtering by
- Genre: Doctoral Dissertation
Description
The main objective of this work is to study novel stochastic modeling applications to cybersecurity aspects across three dimensions: Loss, attack, and detection. First, motivated by recent spatial stochastic models with cyber insurance applications, the first and second moments of the size of a typical cluster of bond percolation on finite graphs are studied. More precisely, having a finite graph where edges are independently open with the same probability $p$ and a vertex $x$ chosen uniformly at random, the goal is to find the first and second moments of the number of vertices in the cluster of open edges containing $x$. Exact expressions for the first and second moments of the size distribution of a bond percolation cluster on essential building blocks of hybrid graphs: the ring, the path, the random star, and regular graphs are derived. Upper bounds for the moments are obtained by using a coupling argument to compare the percolation model with branching processes when the graph is the random rooted tree with a given offspring distribution and a given finite radius. Second, the Petri Net modeling framework for performance analysis is well established; extensions provide enough flexibility to examine the behavior of a permissioned blockchain platform in the context of an ongoing cyberattack via simulation. The relationship between system performance and cyberattack configuration is analyzed. The simulations vary the blockchain's parameters and network structure, revealing the factors that contribute positively or negatively to a Sybil attack through the performance impact of the system. Lastly, the denoising diffusion probabilistic models (DDPM) ability for synthetic tabular data augmentation is studied. DDPMs surpass generative adversarial networks in improving computer vision classification tasks and image generation, for example, stable diffusion. Recent research and open-source implementations point to a strong quality of synthetic tabular data generation for classification and regression tasks. Unfortunately, the present state of literature concerning tabular data augmentation with DDPM for classification is lacking. Further, cyber datasets commonly have highly unbalanced distributions complicating training. Synthetic tabular data augmentation is investigated with cyber datasets and performance of well-known metrics in machine learning classification tasks improve with augmentation and balancing.
ContributorsLa Salle, Axel (Author) / Lanchier, Nicolas (Thesis advisor) / Jevtic, Petar (Thesis advisor) / Motsch, Sebastien (Committee member) / Boscovic, Dragan (Committee member) / Platte, Rodrigo (Committee member) / Arizona State University (Publisher)
Created2023
Description
This dissertation focused on the implementation of urine diversion systems in commercial and institutional buildings in the United States with a focus on control of the urea hydrolysis reaction. Urine diversion is the process by which urine is separately collected at the source in order to realize system benefits, including water conservation, nutrient recovery, and pharmaceutical removal. Urine diversion systems depend greatly on the functionality of nonwater urinals and urine diverting toilets, which are needed to collect undiluted urine. However, the urea hydrolysis reaction creates conditions that lead to precipitation in the fixtures due to the increase in pH from 6 to 9 as ammonia and bicarbonate are produced. Chapter 2 and Chapter 3 describes the creation and use of a cyber-physical system (CPS) to monitor and control urea hydrolysis in the urinal testbed. Two control logics were used to control urea hydrolysis in realistic restroom conditions. In the experiments, acid was added to inhibit urea hydrolysis during periods of high and low building occupancy. These results were able to show that acid should be added based on the restroom use in order to efficiently inhibit urea hydrolysis.
Chapter 4 advanced the results from Chapter 3 by testing the acid addition control logics in a real restroom with the urinal-on-wheels. The results showed that adding acid during periods of high building occupancy equated to the least amount of acid added and allowed for urea hydrolysis inhibition. This study also analyzed the bacterial communities of the collected urine and found that acid addition changed the structure of the bacterial communities.
Chapter 5 showed an example of the capabilities of a CPS when implemented in CI buildings. The study used data mining methods to predict chlorine residuals in premise plumbing in a CI green building. The results showed that advance modeling methods were able to model the system better than traditional methods. These results show that CPS technology can be used to illuminate systems and can provide information needed to understand conditions within CI buildings.
ContributorsSaetta, Daniella (Author) / Boyer, Treavor H (Thesis advisor) / Hamilton, Kerry (Committee member) / Ross, Heather M. (Committee member) / Boscovic, Dragan (Committee member) / Arizona State University (Publisher)
Created2021
Description
Rapid increases in the installed amounts of Distributed Energy Resources are forcing a paradigm shift to guarantee stability, security, and economics of power distribution systems. This dissertation explores these challenges and proposes solutions to enable higher penetrations of grid-edge devices. The thesis shows that integrating Graph Signal Processing with State Estimation formulation allows accurate estimation of voltage phasors for radial feeders under low-observability conditions using traditional measurements. Furthermore, the Optimal Power Flow formulation presented in this work can reduce the solution time of a bus injection-based convex relaxation formulation, as shown through numerical results. The enhanced real-time knowledge of the system state is leveraged to develop new approaches to cyber-security of a transactive energy market by introducing a blockchain-based Electron Volt Exchange framework that includes a distributed protocol for pricing and scheduling prosumers' production/consumption while keeping constraints and bids private. The distributed algorithm prevents power theft and false data injection by comparing prosumers' reported power exchanges to models of expected power exchanges using measurements from grid sensors to estimate system state. Necessary hardware security is described and integrated into underlying grid-edge devices to verify the provenance of messages to and from these devices. These preventive measures for securing energy transactions are accompanied by additional mitigation measures to maintain voltage stability in inverter-dominated networks by expressing local control actions through Lyapunov analysis to mitigate cyber-attack and generation intermittency effects. The proposed formulation is applicable as long as the Volt-Var and Volt-Watt curves of the inverters can be represented as Lipschitz constants. Simulation results demonstrate how smart inverters can mitigate voltage oscillations throughout the distribution network. Approaches are rigorously explored and validated using a combination of real distribution networks and synthetic test cases. Finally, to overcome the scarcity of real data to test distribution systems algorithms a framework is introduced to generate synthetic distribution feeders mapped to real geospatial topologies using available OpenStreetMap data. The methods illustrate how to create synthetic feeders across the entire ZIP Code, with minimal input data for any location. These stackable scientific findings conclude with a brief discussion of physical deployment opportunities to accelerate grid modernization efforts.
ContributorsSaha, Shammya Shananda (Author) / Johnson, Nathan (Thesis advisor) / Scaglione, Anna (Thesis advisor) / Arnold, Daniel (Committee member) / Boscovic, Dragan (Committee member) / Arizona State University (Publisher)
Created2021