Análisis comparativo de distribuciones estadísticas para el cálculo de una variable hidrológica
The main objectives of this dissertation work, consisted in performing a comparative analysis of non-conventional probability distributions to estimate a hydrological variable, and establish the distributions applied in other knowledge fields with better results, according to the proposed methodolog...
- Autores:
-
Barbosa Hernández, Wilmar Hernán
- Tipo de recurso:
- Fecha de publicación:
- 2020
- Institución:
- Escuela Colombiana de Ingeniería Julio Garavito
- Repositorio:
- Repositorio Institucional ECI
- Idioma:
- spa
- OAI Identifier:
- oai:repositorio.escuelaing.edu.co:001/1250
- Acceso en línea:
- https://catalogo.escuelaing.edu.co/cgi-bin/koha/opac-detail.pl?biblionumber=22442
https://repositorio.escuelaing.edu.co/handle/001/1250
- Palabra clave:
- Distribuciones estadísticas
análisis de frecuencias
bondad de ajuste
hidrología
periodo de retorno
precipitaciones totales anuales
precipitaciones máximas en 24 horas
Probability distributions
Frequency analysis
Best Fit
Hydrology
Return period
Anual rainfall
Maximum 1 day rainfall
- Rights
- openAccess
- License
- Derechos Reservados - Escuela Colombiana de Ingeniería Julio Garavito
| Summary: | The main objectives of this dissertation work, consisted in performing a comparative analysis of non-conventional probability distributions to estimate a hydrological variable, and establish the distributions applied in other knowledge fields with better results, according to the proposed methodology for using in hydrological variables. In general terms the applied methodology consisted in selecting non-record rain gauges with 30-year records of annual rainfall or maximum 1-day rainfall. Missing data was completed and time series of annual rainfall or maximum 1-day rainfall were created. Kolmogorov Smirnov test were carried out for detecting normality in each time series. If time series fitted to normal distribution, parametric test were applied. In contrast, if the time series follow a free distribution, nonparametric test were applied. Finally, homogeneity and stationarity test were carried out. if all test were accepted, time series of selected rain gauges, were fitted to different probability distributions through EasyFit 5.6 software, and score ranking from 5 to 1 for each time series was set out, where 5 was assigned to the best goodness of fit and 1 to the probability distribution with the fifth best goodness of fit. Finally, scores were added and then analysis result were carried out. This research let to establish non-conventional probability distributions with best fitting. Best score in the ranking was achieved by log logistic 3P for anual rainfall, and generalized extreme value, Burr and Johnson SB for maximum 1-day rainfall. |
|---|