Flood Prediction Using Multi-Layer Artificial Neural Network in Monitoring System with Rain Gauge, Water Level, Soil Moisture Sensors

Flood is one of the most destructive natural phenomena that happens on most part of the world. Notably in the Philippines, this was a major issue as it can lead to damage of properties, damage to infrastructures or even loss of lives. Current systems adhere to solve issues to prevent devastating disasters caused by floods. In this study, a system is developed to predict flood level based on real-time monitoring sensors and systems. The system predicts in advance the flood level based on the current data it gathered from sensors integrated in a real-time monitoring system. Multi-layered artificial neural network with the aid of MATLAB was used in the development of the prediction model. In the training, test, validation and overall dataset, the network showed a very good goodness-of-fit specifically 0.99889 for the training dataset, 0.99362 for the test data set, 0.99764 for the validation dataset and 0.99795 considering all the data in the dataset. The network was then programmed and integrated in the system in the actual setup. The model is validated by running trials with certain inputs and predicted flood level as the output and is compared to the actual flood level after a certain period of time. The flood prediction system showed an RMSD value of 2.2648 which signifies a small overall difference between the predicted flood level and actual flood level across the whole dataset tested in the actual setup.