Assessment of the capacity credit of wind power in Mexico Juan Pablo Y  a ~ nez 1 , Alexander Kunith 2 , Roberto Ch  avez-Arroyo, Alejandro Romo-Perea, Oliver Probst * Physics Department, Tecnol ogico de Monterrey, Campus Monterrey, Av. Eugenio Garza Sada 2501, Monterrey CP 64849, NL, Mexico article info Article history: Received 17 August 2012 Accepted 24 June 2014 Available online 12 July 2014 Keywords: Capacity credit Wind power Loss of load expectation Geographic information systems North American regional reanalysis Regional diversification abstract A comprehensive assessment of the capacity credit of potential wind power developments in Mexico has been conducted for the first time. The analysis is based on an 80 m wind speed map generated from the North American Regional Reanalysis (NARR) data base and a set of restrictions, including proximity to transmission lines and major roads. Potential wind farm sites complying with all restrictions were populated with wind farms according to different scenarios; consecutive deployment of wind power from 1% to 15% system penetration was considered in all cases. In a set of one-region scenarios the evolution of the capacity credit was studied for different levels of intra-regional diversification. Near- generic decay according to a power law was observed at high penetration levels, whereas a notorious benefit was obtained from diversification at low and intermediate wind power penetration. In order to assess the potential benefits of inter-regional diversification, an optimization procedure was conducted. A significant improvement of the capacity credit decay curve was obtained for all levels of penetration. Optimal sets are characterized by a balanced utilization among regions with a relative insensitivity with respect to the exact composition of the wind farm set. The results are believed to be useful for the expansion planning of the Mexican electric grid. © 2014 Elsevier Ltd. All rights reserved. 1. Introduction As the participation of fluctuating renewable energy sources in the power sector grows, the assessment of the impact of renewable power plants on the reliability of the electric grid becomes more important [1e3]. While traditionally it has been assumed that fluctuating generating facilities do not contribute to the reliability of the grid in terms of an effective firm capacity, this perception has been changing over the last decade, leading to the general recog- nition of a capacity credit of renewable energy sources [4,6,8]. Different indices such as the Loss of Load Expectation (LOLE) [5e7] and the Loss of Load Probability (LOLP) [6,7] have been traditionally used to discuss system reliability on transmission networks and can be conveniently generalized to include the effect of fluctuating power sources. The different approaches to calculating capacity credit of renewable energy sources have been reviewed by different authors, including Milligan and Porter [6] and Dent and al [8]. Rather than qualitatively different from conventional generating plants such as coal- or gas-fired plants based on thermal cycles renewable energy power plants differ only quantitatively from the former in the sense that the variable characterizing availability of conventional plants has to be replaced by the actual power output of the renewable power plant. Evidently, the availability of a con- ventional plant is generally much higher (of the order of 80%e90%) than the average power output of a esay e wind farm (with a typical capacity factor of the order of 30%e40%) and the fluctua- tions of the output of the conventional plant are much lower than those of the renewable plant, but the conceptual framework to be used is the same. The general approach with regards to the contributions of a renewable energy source to system reliability is to determine the effective firm capacity that can be ascribed to the fluctuating source. A typical procedure [9], the one used in the present work, is to first determine the Loss of Load Expectation (LOLE) of the system without the addition of the renewable energy capacity, i.e. the number of yearly hours during which the available generating ca- pacity is unable to meet the load. The effective firm capacity of the projected renewable nameplate generation capacity P 0 is then calculated by either (i) assuming that a certain amount of firm ca- pacity (i.e. capacity assumed to have an availability of 100%) is replaced by P 0 , while maintaining the system LOLE at its original * Corresponding author. E-mail addresses: oprobst@itesm.mx, oliver.probst@gmail.com (O. Probst). 1 Present address: DESY, D-15735 Zeuthen, Germany. 2 Present address: Technical University of Berlin, Berlin, Germany. Contents lists available at ScienceDirect Renewable Energy journal homepage: www.elsevier.com/locate/renene http://dx.doi.org/10.1016/j.renene.2014.06.038 0960-1481/© 2014 Elsevier Ltd. All rights reserved. Renewable Energy 72 (2014) 62e78