Raja Mohamad, Raja Nurul Mardhiah and Wan Ishak, Wan Hussain (2018) Forecasting reservoir water level based on the change in rainfall pattern using neural network. In: 2nd Conference on Technology & Operations Management (2ndCTOM), February 26-27, 2018, Universiti Utara Malaysia, Kedah, Malaysia. (Unpublished)

PDF Restricted to Registered users only Download (623kB) | Request a copy

Abstract

Reservoir water level is a level of storage space for water.During heavy rainfall, waterstorage space is used to hold excessive amount of water. During less rainfall, water storage maintains the water supply for its major uses.The change in rainfall pattern may influence the water storage. Thus, understanding the change in rainfall pattern can be used in gate opening decision.On top of that, the upstream precipitation is always not coincide with the consequences and the flood downstream usually cause expensive damages and great devastation as well.This study focus on the analysis of the upstream rainfall data in order to obtain the rainfall pattern.This study deployed basic steps in Artificial Neural Network (ANN) modeling which are data selection, data preparation, data pre-processing and finally Neural Network model development and evaluation. The performance of ANN was based on MAE (mean absolute error) and RMSE (root mean square error). In this study, three datasets have been formed that represent the change in upstream rainfall pattern. The findings show that the best RMSE achieve is 0.644 from the third dataset.

Item Type:	Conference or Workshop Item (Paper)
Additional Information:	Organized by: School of Technology Management and Logistics (STML), Universiti Utara Malaysia
Uncontrolled Keywords:	reservoir water level, rainfall pattern, artificial neural network.
Subjects:	Q Science > QA Mathematics > QA76 Computer software
Divisions:	School of Technology Management & Logistics
Depositing User:	Mrs. Norazmilah Yaakub
Date Deposited:	04 Jun 2018 01:11
Last Modified:	04 Jun 2018 01:11
URI:	https://repo.uum.edu.my/id/eprint/24238

Actions (login required)

View Item