Performance of artificial neural network and regression techniques for simulation model in reservoir inter-relationships

Abstract

In fact, the reservoir simulation is an essential step while developing the optimal operation policy for dam and reservoir. Generally, an accurate simulation for the reservoir characteristics should lead more reliable and robust optimization model for certain reservoir. The major challenge in the reservoir simulation is the non-linearity behavior of inter-relationships between the reservoir elevation, surface area and storage. The existing traditional modeling methods usually solve such problem in linear fashion, thus, it achieves relatively poor accurate simulation especially at extreme values, which affect the reliability of the optimization model. The use of artificial neural networks (ANNs) is becoming increasingly common in the analysis of hydrology and water resources problems and it may replace the classical regression model hitherto used most especially for non-linear system. Two different types of ANN namely, feed-forward back-propagation neural network (FBNN) and radial basis function neural network (RBFNN) were used in this study to simulate the inter-relationships between elevation, surface area and storage capacity at Langat reservoir system, Malaysia. In addition, classical auto regression (AR) model was developed for comparative analysis over the proposed ANN model. The main finding of this study showed that proposed ANN model could significantly improve the simulation accuracy over the classical AR mode. On the other hand, the results obtained for RBFNN were found to be more accurate than the simulation of AR and FBNN. This study thus concludes that the ANN method is more suitable to simulate the reservoir behavior than the classical regression model. Key words: Artificial neural network, elevation-storage, elevation-surface area, reservoir simulation.