OSG has organized the following events:
|
Generation Scheduling under Renewable-Based Uncertainty via Two-Stage Adaptive Robust Optimization, April 19, 2018
José M. Arroyo Universidad de Castilla-La Mancha Within the context of current day-ahead electricity markets, the presentation examines the use of two-stage adaptive robust optimization as a relevant tool to handle renewable-based uncertainty in generation scheduling. Unlike alternative approaches to deal with uncertainty, neither accurate probabilistic information nor a discrete set of uncertainty realizations are required. Rather, uncertainty is modeled by decision variables within a deterministic uncertainty set. Hence, the size of the robust models does not depend on the dimension of the space of uncertainty realizations belonging to the uncertainty set, thereby providing a computationally efficient framework. In addition, an easy control of the degree of conservativeness can be implemented. The resulting robust counterparts are instances of mixed-integer trilevel programming. Practical modeling aspects allow using effective decomposition-based techniques that guarantee finite convergence to optimality. Results from several case studies illustrate the effectiveness of the two-stage robust setting. |
| ||
|
Revenue and Network-Constrained Day-Ahead Market Clearing under Marginal Pricing, April 12, 2018
Natalia Alguacil Universidad de Castilla-La Mancha In this seminar a practical day-ahead auction model, where generation revenue constraints are explicitly incorporated in the problem formulation, will be presented. The revenue-constrained market-clearing procedure includes the effect of the transmission network, inter-temporal constraints associated with generation scheduling, demand-side bidding, and marginal pricing. This auction design is an instance of price-based market clearing which features two major complicating factors. First, locational marginal prices become decision variables of the optimization process. In addition, producer revenues are formulated as bilinear and nonconvex products of power outputs and market-clearing prices. The resulting problem is formulated as a mixed-integer nonlinear bilevel program with bilinear terms for which available solution techniques rely on heuristics, approximations, or modeling simplifications. This work shows a novel and exact methodology whereby the original problem is recast as an equivalent single-level mixed-integer linear program. As a consequence, finite convergence to optimality is guaranteed and the use of standard commercial software is allowed. The proposed transformation is based on duality theory of linear programming, Karush-Kuhn-Tucker optimality conditions, and integer algebra results. |
| ||
|
Increasing Electric Vehicle Adoption Via Strategic Siting of Charging Stations, March 28 2018
Martim Joyce-Moniz GERAD In order to increase electric vehicle (EV) adoption it is necessary for a central authority to drive the initial investment in charging infrastructures. Recent works on optimization problems for siting and sizing of charging stations for EVs have started addressing this issue by considering strategic multi-period optimization problems. One limitation of these works, however, is that they consider the demand (i.e. number of EVs and their geographical distribution) over time to be static and given as an input. In this presentation, we present a more holistic optimization framework that considers how new infrastructure impacts EV demand growth, and how the infrastructure can be installed in a way that it properly responds to future demand. This framework has been validated by Hydro-Québec and is being applied to help achieve Quebec's objective of having 100,000 EVs on the roads by 2020, and 1,000,000 by 2030. |
| ||
|
On the packing process in a shoes manufacturer, March 22 2018
Manuel V.C. Vieira Nova University of Lisbon In this presentation we describe a real life application of a container loading problem for a children’s shoes manufacturer. First, we choose the appropriate shoe box for each pair of shoes depending on the model and size before production starts. Then, the shoe boxes are packed in several cardboard boxes with variable dimensions. We present an integer programming model that chooses the shoe boxes for each model size, and a MINLP model which decides how to pack the shoe boxes and the size of cardboard boxes. This container loading problem is classified as open dimension problem, with three open dimensions. We approximate the MINLP model with a MILP and we compare it by using BARON on the former model and we run CPLEX on the latter model. |
| ||
|
Assessing Wind Resource Adequacy for Peak Demand, March 15 2018
Kristen R. Schell University of Michigan There are almost 50,000 individual wind turbines currently operating in the continental United States. Federal tax credits, as well as local price differences, have helped spur past and future investment in the wind industry. While this increase in wind capacity, at times, provides a remarkable percentage of load with renewable generation, variable wind power output is not often synchronous with peak demand, which raises the issue of its contribution to overall resource adequacy. We propose a new method for assessing wind resource adequacy in the planning phase, utilizing cross-spectra analysis of wind speed and electricity system load time series. The results indicate which geographic locations in an electricity system have wind resource potential that is most able to contribute to meeting peak load. This metric gives wind farm planners information on where to site wind farms that reduce reliability risk and increase supply adequacy. Such information is particularly important as electricity systems move toward maximum levels of variable renewable power penetration. Results are shown for three major electricity markets of interest – CAISO in California, NYISO in New York and ERCOT in Texas. |
| ||
|
Functional Econometrics of Multi-Unit Auctions: An Application to the New
York Electricity Market, March 01 2018
David Benatia Université de Montréal This paper proposes a novel approach for the empirical analysis of multi-unit auctions, to which participants submit supply or demand functions observable by the econometrician. The approach allows for the evaluation of firm-level market power in a private information setting, and avoids having to model the market mechanism. It relies on econometric methods that treat the observed bid functions as function-valued random elements. Notably, a functional instrumental variable estimator is developed. The method is applied to the New York electricity market using rich data on firm-level bids and marginal costs for 2013-2015. In this market, daily bids are disclosed three months later in order to limit strategic behaviors. I estimate firm-level market power and compare actual bidding behavior to profit-maximizing behavior under private information. I find consistent evidence of strategic bidding, suggesting that firms are well aware of their own market power and behave accordingly. Therefore, the late disclosure of bids is not sufficient to preclude firms from acting strategically, most likely due to the repeated nature of those auctions. |
| ||
|
Regulatory, Rate, and Market Design for Energy Storage , November 1st 2017
Professor Ramteen Sioshansi Ohio State University Energy storage is a unique grid asset, in that it can provide services that are functionally similar to those provided by generation, transmission, and distribution assets. Most of our market designs and regulatory constructs assume that assets primarily fall into one of these three categories. This talk will introduce the challenges that are raised by this dichotomy between the capabilities of energy storage and the regulatory and market treatment of grid assets. It will also discuss some potential solutions to overcoming these challenges, including the use of storagecapacity rights. |
| ||
