Streamflow forecast model using ANN

Abstract

Assam is one of the highly flood-affected states in India, mostly by river Brahmaputra and its territories. Millions of people become homeless, jobless and face a life threat, thousands lose their lives during floods. This article presents a flood monitoring and warning system using artificial neural networks (ANN) and embedded systems. Sensor nodes communicate with a cloud server to transmit real-time data, aiding in flood prediction. An ANN, trained on a decade of data from Sonitpur district in Assam, predicts flood probabilities. The ANN accurately mimics rainfall data with 97-98% accuracy by considering temperature, humidity, and rainfall. It correlates these factors to determine water levels and predict flood probabilities. The system monitors environmental conditions and flood likelihood based on past data, displaying real-time information on temperature, humidity, rainfall, and water levels.

KEYWORDS:

• ANN
• flood monitoring
• wireless sensor networks
• Assam flood probability
• natural hazards

Disclosure statement

No potential conflict of interest was reported by the author(s).