Developing water losses and reservoir final storage forecast has become an increasingly important task for reservoir operation. Accurate forecasts would lead to better monitoring of water quality and more efficient reservoir operation. Therefore, the flash flood and water crisis problems in Malaysia can be reduced. Artificial neural networks (ANN) models with radial basis function (RBF) have been determined for high efficiency and accuracy, especially in the dynamics system. In this study, the proposed ANN Prediction Model is being developed by using inflow, the release of dam, initial and final storage of the reservoir as input, whereas the water losses from the reservoir as output. All the data collected over 11 years (1997–2007) at Klang Gate reservoir has been used to develop and test model output. The results indicated that the proposed model could provide monthly forecasting with maximum root mean square error of ± 20.07%. The advantages of this ANN model are to provide information for water losses, final storage, and variation of water level for better reservoir operation.

开发水损失和水库最终储水量预测已成为水库运行中越来越重要的任务。准确的预测将导致对水质的更好监测和更有效的水库运行。因此，可以减少马来西亚的山洪和水危机问题。已经确定具有径向基函数（RBF）的人工神经网络（ANN）模型，以提高效率和准确性，尤其是在动力学系统中。在这项研究中，拟议的ANN预测模型是通过使用流入量，水坝的释放，水库的初始和最终存储作为输入，而水库中的水损失作为输出而开发的。在巴生门水库11年（1997-2007年）中收集的所有数据已用于开发和测试模型输出。结果表明，所提出的模型可以提供最大的均方根误差为±20.07％的月度预测。该人工神经网络模型的优点是可以提供有关失水量，最终储水量和水位变化的信息，以实现更好的水库运行。