Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

MC42

average), but such strategies cannot apply to fleet managed SAEVs (continually in-service). This research proposes a SAEV SC framework to shift electricity demand away from peak use hours (price-based SC) or towards hours with high renewable generation (generation-based SC). A case study from Puget Sound region integrates the regional travel demand model, real-time energy prices, and regional renewable generation data. 4 - Statistical Approaches to Large-scale Building Modeling and Grid Applications Eric Wilson, National Renewable Energy Laboratory, Golden, CO, United States This presentation will provide an overview of the data sources and methodology behind ResStock, a highly granular bottom-up engineering model that represents national, regional, and local housing stocks using hundreds of thousands of sub- hourly building energy models. Several real-world applications of ResStock will be discussed, including how ResStock is being used by cities to answer their most pressing questions about electrification, resilience, and achieving 100% renewable energy goals. n MC44 North Bldg 227C Time-Varying Optimization and Learning in Power Systems Sponsored: Energy, Natural Res & the Environment/Electricity Sponsored Session Chair: Steven Low, California Institute of Technology, Pasadena, CA, United States Co-Chair: Yujie Tang, California Institute of Technology, Pasadena, CA, 91125, United States 1 - Feedback-based Power System Optimization with Time-varying Constraints Adrian Robert Hauswirth, ETH Zurich, Physikstrasse 3, Zurich, 8055, Switzerland, Saverio Bolognani, Gabriela Hug, Florian D÷rfler We study continuous-time, non-smooth dynamical systems which arise in the context of time-varying non-convex optimization, as for example the feedback- based optimization of power systems. We generalize the notion of projected dynamical systems to time-varying domains and derive conditions for the existence of solutions. These requirements relate to the time-varying nature and constraint qualifications for the optimization problem as well as the structure of the feedback control. To illustrate the necessity and usefulness of such a framework, we consider a simple yet insightful power system example, and discuss the implications of the proposed conditions for feedback optimization schemes. 2 - Asynchronous and Distributed Tracking of Time-varying Fixed Points Andrey Bernstein, Senior Scientist, National Renewable Energy Laboratory, Golden, CO, United States We develop an algorithmic framework for tracking fixed points of time-varying contraction mappings. Analytical results for the tracking error are established for the cases where: (i) the underlying map changes at each step of the algorithm; (ii) only an imperfect information of the map is available; and, (iii) the algorithm is implemented in a distributed fashion, with communication delays and packet drops leading to asynchronous algorithmic updates. The analytical results are applicable to several classes of problems, including time-varying contraction mappings emerging from online and asynchronous implementations of gradient- based methods for time-varying convex programs. 3 - An Optimal and Distributed Feedback Voltage Control under Limited Reactive Power Guannan Qu, Harvard University, 33 Oxford Street, Room 336, Cambridge, MA, 02138, United States, Na Li In this talk, we propose a distributed voltage control in power distribution networks through reactive power compensation. The proposed control can (i) operate in a distributed fashion where each bus makes its decision based on local voltage measurements and communication with neighbors, (ii) always satisfy the reactive power capacity constraint, (iii) drive the voltage magnitude into an acceptable range, and (iv) minimize an operational cost. We will also present the underlying methodology behind the voltage control, a variant of primal-dual gradient algorithm (dynamics) that have geometric convergence rate (exponential stability) guarantee, which may be of independent interest.

n MC42 North Bldg 227A Simulation Optimization in Networks Sponsored: Simulation Sponsored Session Chair: Kalyani Nagaraj, Oklahoma State

1 - Rare-event Simulation for Electric Power Distribution Networks Chang-Han Rhee, Centrum Wiskunde and Informatica, Jakoba Mulderplein 164, Amsterdam, 1018 MZ, Netherlands, Bert Zwart, Niek Vasmel We investigate a rare-event simulation problem for the electric power distribution network. We reformulate the non-linear problem described by the power flow equation into a simulation problem of a Markov chain. Building on this formulation, we propose variance reduction methods for the simulation estimation of voltage drop probability. 2 - The Population Dynamics Algorithm Mariana Olvera-Cravioto, University of California, Berkeley, 4125 Etcheverry Hall, Berkeley, CA, 94720, United States This talk will focus on the convergence of the population dynamics algorithm, which produces sample pools of random variables having a distribution that closely approximates that of the special endogenous solution to a variety of branching stochastic fixed-point equations often encountered in the analysis of random graphs. Specifically, we show its convergence in the Wasserstein metric of order p (p >= 1) and prove the consistency of estimators based on the sample pool produced by the algorithm. 3 - Risk-averse Set Covering Problems Hao-Hsiang Wu, University of Washington, 5000 25th Avenue NE, Seattle, WA, 98105, United States, Simge Kucukyavuz We consider probabilistic set covering problems under conditional value-at-risk. Suppose that we have an oracle that computes the risk efficiently for a given solution. Using this oracle, we propose methods for solving the risk-averse set covering problem exactly. We give valid inequalities that strengthen the formulation. We report our computational experience with the proposed methods on a probabilistic set covering problem that admits an efficient risk oracle. n MC43 North Bldg 227B Energy and Climate 8: Urban Energy Systems Modeling, Control, and Grid Interactions Emerging Topic: Energy and Climate Emerging Topic Session Chair: Thomas Deetjen, Carnegie Mellon University,Austin, TX, 78701, United States 1 - Using Optimization to Identify Decarbonization Pathways for Cities and Districts Ashreeta Prasanna, Empa Swiss Federal Laboratories Decarbonization of the energy supply system of a city is an extremely complex process. Planning for centralized and distributed energy provision over a longer time frame involves consideration of the objectives of several stakeholders, potential technology improvement over time, impact of local and national policy, and age of the existing building and technology stock. Using a case study, we show how all these aspects can be considered within a multi-stage optimization model to identify optimal pathways for decarbonization of a city. 2 - ScoutùSoftware for National Building Energy Efficiency Impact Assessment Jared Langevin, Lawrence Berkeley National Laboratory Scout is an open-source software program developed by the U.S. Department of Energy’s Building Technologies Office (BTO) that estimates the energy, carbon, and operating cost impacts of energy conservation measures (ECMs) on U.S. residential and commercial buildings across both long- and short-term time horizons. In this work, Scout’s core analysis capabilities are described, including building technology stock-and-flow dynamics and the adoption logic used to compete ECMs that apply to the same segments of baseline energy use. Recent extensions of Scout to time-sensitive valuations of energy efficiency are also discussed. 3 - Smart Charging Management of Shared Autonomous Electric Vehicles and Implications for the Grid T. Donna Chen, Assistant Professor, University of Virginia, P.O. Box 400742, Charlottesville, VA, 22904, United States, Zhuoyi Zhang As shared autonomous electric vehicle (SAEV) fleets roll out to the market, fleet charging will significantly impact energy demand. Existing EV smart charging (SC) research focus on privately-owned EVs (utilized for 5% of the day, on

200