2015 Informs Annual Meeting

MD57

INFORMS Philadelphia – 2015

MD57 57-Room 109B, CC Planning Models in Electric Power Systems Sponsor: ENRE – Energy I – Electricity Sponsored Session Chair: Anya Castillo, Federal Energy Regulatory Commission, 888 1st Street NE, Washington DC, United States of America, anya.castillo@ferc.gov 1 - Optimal Portfolio Investment and Coordinated Scheduling of an Energy Storage Merchant in the Energy Roderick Go, Johns Hopkins University, 3400 N. Charles St., Ames Hall 313, Baltimore, MD, 21218, United States of America, rgo1@jhu.edu, Anya Castillo, Dennice F. Gayme, Sonja Wogrin We assess strategic behavior of a merchant energy storage provider in the bulk power market through a bilevel model to represent sequential decisions in investment and operations. We model optimal portfolio investments based on siting, sizing, and technology mix, and explore the effect of strategic operations, such as coordinated scheduling, on decisions. We transform this model into a math program with equilibrium constraints (MPEC), and approximate and solve as a mixed integer program (MIP). 2 - Proactive Transmission Planning: A Case Study of the Eastern Interconnection Evangelia Spyrou, PhD Student, Johns Hopkins University, 3400 N Charles Street, Dept of Geography, Johns Hopkins University, Baltimore, MD, 21218, United States of America, elina.spirou@gmail.com, Benjamin Hobbs, Jonathan Ho, Randell Johnson, James Mc Calley Traditional transmission planning procedures are being challenged by renewable integration due to their reactive character. Meanwhile academic literature proposes the concept of proactive transmission planning. A mixed integer linear program is applied to estimate the benefits of proactively considering response by generation investments to transmission investments. We attempt to examine features of planning procedures that could impede or facilitate optimal planning. 3 - Unit Commitment Approximations in Generation and Transmission Planning: Efficiency & Accuracy Benjamin Hobbs, Professor, The Johns Hopkins University, 3400 North Charles Street, Baltimore, MD, United States of America, bhobbs@jhu.edu, Saamrat Kasina, Jonathan Ho, Sonja Wogrin Alternative tight relaxations of unit commitment problems that enable large planning models to be solved with operating subproblems that capture ramp, start-up, and pmin limits and costs. We examine their performance in the context of generation and transmission expansion models, including a stochastic programming analysis of the western interconnection of North America. 4 - Reserve Determination Methods for Variable Generation Robert Entriken, EPRI, 3420 Hillview Avenue, Palo Alto, CA, United States of America, rentrike@epri.com We present results of a survey of existing practices in certain power system operators for determining operating reserve requirements for system operators faced with growing penetrations of variable renewables, such as wind or solar generators. Building on existing practices, we review methods proposed in planning and integration studies, as well as in academia, which may become useful as renewable penetrations increase. MD58 58-Room 110A, CC Multi-Agent Decision-Making for Smart Grids Operation II Sponsor: ENRE – Energy I – Electricity Sponsored Session Chair: Amin Kargarian, Carnegie Mellon University, 5000 Forbs Ave, Pittuburgh, PA, 15232, United States of America, amin.kargarian@gmail.com 1 - Distributed State Estimation and Energy Management in Smart Grids Soummya Kar, Assistant Research Professor, Carnegie Mellon University, Electrical and Computer Engineering, CMU, Pittuburgh, PA, 15232, United States of America, soummyak@andrew.cmu.edu Generally, it is expected that the grid of the future would differ from the current system by the increased integration of distributed generation, distributed storage, demand response, power electronics, and communications and sensing technologies. In this paper, we discuss distributed approaches, all based on

consensus+innovations, for two common energy management functions: state estimation and economic dispatch. 2 - Computational Look-ahead SCOPF via ADMM Sambuddha Chakrabarti, Graduate Student, UT Austin, 1616 Guadalupe Street, Austin, TX, 78705, United States of America, sambuddha.chakrabarti@gmail.com, Matt Kraning, Ross Baldick, Eric Chu, Stephen Boyd We present computational scheme and results of the ADMM based Proximal Message Passing as applied to solve the look ahead SCOPF to limit post fault line temperature and current to safe values wrt next set of outages. 3 - Adaptive Bidding Strategies of a Load Serving Entity with Distributed Energy Resources Jhi-Young Joo, Assistant Professor, Missouri University of Science and Technology, 301 W. 16th St, 235 Emerson Electric Co. Hall, Rolla, MO, 65409, United States of America, joojh@mst.edu This talk concerns two problems solved by an agent, a load serving entity (LSE), within a large-scale energy system. An LSE with different types of demand and energy resources optimizes energy schedule by mathematical programming. On the other hand, to optimize bids into the markets, a learning algorithm is used to adapt to the uncertain market conditions and rewards. The interdependencies between these two problems within an agent and among multiple agents are examined. 4 - Active Distribution Grid Operation: A System of Systems Framework Amin Kargarian, Carnegie Mellon University, 5000 Forbes Ave, Pittuburgh, PA, 15232, United States of America, amin.kargarian@gmail.com A system of systems framework is presented to operate an active distribution grid composing of several independent entities. The grid structure includes two layers where the distribution company is in upper layer and microgrids are in lower layer. A hierarchical optimization algorithm is presented to optimally operate the entire active distribution grid. MD59 59-Room 110B, CC Forest & Timber Management Sponsor: ENRE – Environment II – Forestry Sponsored Session Chair: Nick Kullman, Masters Student, University of Washington, 360 Bloedel Hall, Seattle, WA, United States of America, nick.kullman@gmail.com 1 - A Joint Model of Strategic Forest Management, Capacity Expansion and Logistics Eldon Gunn, Dalhousie University, Halifax, NS, B3H 4R2 Eldon.Gunn@Dal.Ca This paper presents a mixed integer programming model that enables the integrated analysis of strategic forest management, forest industry capacity and the transport logistics that connect them. Some insights that arise from this model are discussed. 2 - Route Selection in Forest Tansportation Patrik Flisberg, Creative Optimization, Tokai, 7945 Cape Town, South Africa, pafli@mweb.co.za, Mikael Ronnqvist, Gunnar Svenson Determining the best route for logging trucks is difficult as many road features need to be considered. We describe a system called Calibrated Route Finder that is used to invoice about 50% of all 2 million forest transports done annually in Sweden. This system has gradually been developed based on reporting and requests from the users. Recently, we have included detailed description on stops, acceleration and breaking to describe emissions and times. We report on detailed testing and analysis. 3 - Joint Production of Timber and Sitka Deer Habitat Capability on the Tongass National Forest Michael Bevers, Dalhousie University, Halifax, NS, B3H 4R2 beversm@gmail.com, Curt Flather, Yu Wei, Greg Hayward, Mary Friberg, Thomas Hanley, Ben Case The Tongass NF uses the FRESH model to estimate Sitka deer habitat capability measured in deer-days based on digestible dry matter and protein from hundreds of forage species occurring in dozens of vegetative communities potentially affected by timber harvests. Landscape effects on deer are accounted for using a moving window analysis. We developed a MILP formulation incorporating FRESH calculations into a whole-stand timber harvest scheduling model for spatially optimizing joint production of timber and deer habitat capability on management units of the Tongass NF.

250