3RD INTERNATIONAL CONGRESS ON TECHNOLOGY - ENGINEERING & SCIENCE - Kuala Lumpur - Malaysia (2017-02-09)

Modeling The Behavior Of A Reservoir System Employing Performance Indices In Kelantan Catchment
Modeling the duration of the critical period (CP) and therefore developing the exact behavior model of a reservoir system prior to analysis is beneficial because then the input data interval can be selected to match the requirement. In this sense, over-year systems (CP > 1 year) can be sufficiently analyzed applying annual time-series data, whereas for within-year systems (CP < 1 year), only the critical 12 months in the data record are required to be considered in the analysis. The current standard demand parameter method in the literature for prediction of the behavior of reservoirs is based only on demand and coefficient of variation of annual flows. Nevertheless, performance measures such as reliability and vulnerability also play the role in the critical period of reservoir systems. Therefore, the objective of this study is to develop a predictive relationship for critical period involving standard demand parameter and both reliability and vulnerability performance indices as for use during the reservoir planning stage. The Kelantan catchment was chosen as the case study. The reservoir on this catchment was analyzed using 1000 sequences of synthetic data having the same length as historical data involving both time-based reliability and vulnerability performance indices by modified Sequent Peak Algorithm (SPA) for different demands. Modified SPA is capable of involving performance indices in the analysis of the reservoir systems which makes this method superior to the other techniques in the literature that are not able to consider such extra parameters in the reservoir system yield analysis. According to the results of the modified SPA, the new predictive equation for the critical period of both within-year and over-year reservoir systems have been developed employing reliability and vulnerability indices at two stages. In the first stage, the power relationships are developed for the critical period at different reliability and vulnerability indices. In the second stage, the parameters that were obtained in the first stage for the power relationships are modeled using the linear regression relationships applying the reliability and vulnerability indices for the study system using the method of the least squares. At the two stages, the regression analysis produced statistically satisfactory results. Finally, the R2 obtained over the complete range of the critical period prediction is high, being 0.9824 for the study system. Hence, this equation should enable the prediction of the critical period for the study reservoir system efficiently which can be used to determine the precise behavior of the system for selecting the input streamflow data interval in the reservoir system analysis.
Issa Saket Oskoui, Rozi Abdullah