Abstract
Data generation mechanisms have been widely applied in hydrology. Models are built for generation of data having the statistical properties of the historical records. The creation of synthetic time series starts with the generation of independent normal variables with average zero and variance one, then adding the time and spatial dependence structure as well as periodic components, whichever necessary. The data generation can be accomplished via the analysis of the historical data to check its suitability for generation, Selection, identification of the form, estimation of parameters, & check of the data generation model and Application of the model & testing of the results. This paper summarizes the required work to be done as per the above steps taking the autoregressive moving average as an example of the data generation model Keywords: Induced Voltage, – Electric Fields, HVDC Transmission, Finite Element Method, Hybrid Transmission Lines. DOI : 10.7176/ISDE/10-1-02
Highlights
The designers of water resources systems have realized that evaluating their designs using past or historical data provided no guarantee that the design would perform satisfactorily in the future because future flow sequences will not be the same as past flow sequences (Haan, 1982)
Data Generation Models In general, data generation can be accomplished in three steps, Analysis of the historical data to check its suitability for generation, Selection, identification of the form, estimation of parameters, and check of Application of the model and testing of the results
If the periodicity is clearly noticed, it should be removed using both parametric and non-parametric methods as follows, 5.4.1 Parametric Method for Separating Periodicity: To economize on the number of statistics needed for mathematical description of time series, the periodic component can be separated by superimposed harmonics
Summary
The designers of water resources systems have realized that evaluating their designs using past or historical data provided no guarantee that the design would perform satisfactorily in the future because future flow sequences will not be the same as past flow sequences (Haan, 1982). The last statement describes the need for data generation in order to obtain new time series simulating the possible future flows. Synthetic data series are generated by many models using autoregressive (AR) processes, Thomas-Fiering model (1967), method of fragments (Srikanthan and McMohan(1985)) and its modified version (Maheepala and Perera (1996)) , the nonparametric approach model (Sharma and O'Neil (2002))and wavelet approach (Ünal, Aksoy and Akar (2004))
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
