Development of a Lightning Prediction Model Using Machine Learning Algorithm: Survey.

Abstract

This research is aimed at preventing broadcast equipment from lightning damage. Inview of the location in which my broadcast outfit is located (located in a valley; some few meters above sea level in the Confluence of Lokoja Kogi State Nigeria). Several improvement of earthing and installation of lightning arresting facilities, there has not been significant change protecting broadcast equipment from lightning. The solution I proffered is to isolate all electrical connection from equipment. Lightning as a natural phenomenon is very unpredictive and destructive which can occur during transmission. How do we know the day and time distructive lightning will come? The answer is to develop a lightning prediction system that is accurate. When lightning lead time is known, personnel on duty will be alerted to isolate all broadcast equipment from the mains and central earth connection. Since the lightning prediction system has to be localized. Deployment of machine learning algorithm is most appropriate. The use of ten(10) years Weather Numerical Values(2012-2022) such as rainfall, atmospheric pressure, relative humidity, temperature and lightning records which are gotten from Nigeria Meteorological Agency (NIMET), Lokoja Area Office, Kogi State. This parametization are factors that are used to work on the lightning forecast system as lightning will occur at their certain threshold values. The model is intended to be deployed on web application. Prediction model can be localized to support Numerical Weather Prediction in any environment in Nigeria.