Spatio-temporal analysis of extreme wind velocities for infrastructure design = Analisis...

Rexistro bibliográfico

  • Título: Spatio-temporal analysis of extreme wind velocities for infrastructure design = Analisis espacio-temporal de velocidades extremas de viento para diseño de infraestructura
  • Autor: Rodríguez Avellaneda, Alexys Herleym
  • Publicación orixinal: 2020
  • Descrición física: PDF
  • Nota xeral:
    • Colombia
  • Notas de reprodución orixinal: Digitalización realizada por la Biblioteca Virtual del Banco de la República (Colombia)
  • Notas:
    • Resumo: Abstract: This research aims to create non-hurricane non-tornadic maps of extreme wind speeds for the mean recurrence intervals MRIs 700, 1700, and 3000 years, covering the Colombian territory. For infrastructure design, these maps are combined with existing hurricane wind speed studies, to be used as input loads due to wind. For each station with non-thunderstorm wind speeds time histories in the input data, following (Pintar, Simiu, Lombardo, & Levitan, 2015), extreme wind speeds corresponding to each MRI are calculated using a Peaks Over Threshold Poisson Process POT-PP extreme value model, then wind velocities with the same MRI are spatially interpolated to generate continuous maps for the whole study area. The annual exceedance probability for all velocity values in 700, 1700 and 3000 years MRIs output maps are respectively 1/700, 1/1700 and 1/3000. Regarding input data, not only time series of field measurements from IDEAM methodological stations are used, but also postprocessed information coming from the Integrated Surface Database ISD, and ERA5 forecast reanalysis data. This condition demanded a comparison of the different data sources, in order to verify the feasibility in the use of ISD and ERA5, this is downscaling support. The result of the comparison showed little similarity between the different sources, but taking into account that complete and adequate measured data from IDEAM was not available. Before to apply POT-PP, ISD and IDEAM data sources were standardized to meet the requirement of three seconds (3-s) wind gust speed, ten (10) meters anemometer height, and terrain open space condition. Due to the limitation in the classification of thunderstorm and non-thunderstorm data, it was not possible to take real advantage of POT-PP method, which was limited/restricted from non-homogeneous to homogeneous and from nonstationary to stationary, being equivalent to use the most common POT - generalized Pareto approach. Non-hurricane maps were created for data sources ISD and ERA5, using Kriging as spatial interpolation method. After the integration with previous hurricane studies, the results of ERA5 showed the most reliable final maps, despite limitations in the input data to guarantee downscaling support. ISD final map showed very high wind values, which are unlikely. These shortcomings may be corrected when complete IDEAM data-source and storm data classification are available. A complete R tool was implemented to solve the whole process, which is based in copyrighted code for declustering and thresholding generously given by Dr Adam L. Pintar adam.pin tar@nist.gov - NIST, U.S Department of Commerce. Resumen: Esta investigación tiene como objetivo la creación de mapas de velocidades de viento extremo (no-huracán, no tornado) para los periodos de retorno 700, 1700 y 3000 años, cubriendo todo el territorio colombiano. Estos mapas se combinaron con estudios sobre velocidades de viento de huracanes para ser usados como cargas de viento para diseño de infraestructura. Para cada estación en los datos de entrada con historias de tiempo de velocidades de viento tipo ""no tormenta"", siguiendo (Pintar, Simiu, Lombardo, & Levitan, 2015), se calcularon velocidades extremas de viento correspondientes a cada periodo de retorno, usando un modelo de valores extremos de picos sobre el umbral - proceso de Poisson (POT-PP), para posteriormente interpolar velocidades de viento con el mismo periodo de retorno y generar mapas continuos para toda el área de estudio. La probabilidad de excedencia anual para todos los mapas resultantes con valores de velocidad de 700, 1700 y 3000 años, son 1/700, 1/1700 y 1/3000 respectivamente. En relación con los datos de entrada, no solamente se usaron las series de tiempo de los datos de campo del IDEAM, pero también datos de la base de datos ISD (Integrated Surface Database) y datos de reanálisis de ERA5, sin embargo, los resultados finales se basaron solamente en ERA5.
    • © Derechos reservados del autor
    • Colfuturo
  • Forma/Xénero: Tese
  • Idioma: castellano
  • Institución orixe: Biblioteca Virtual del Banco de la República
  • Encabezamento de materia:

Web semántica