Forecasting Photovoltaic Power Generation via an IoT Network Using Nonlinear Autoregressive Neural Network

This research work is an attempt to introduce modern computing techniques as a potential decision-making tool in the field of renewable energy supply and management. We aim to demystify and take advantage of the concept of neural networks to predict the conversion of solar energy by a photovoltaic unit. In order to do so, a smart meter will be built and connected to a low power photovoltaic panel, the smart meter once in operation will capture a set of data, send them autonomously to a remote server over a LoRa IoT network which will be then processed to make predictions about the amount of power produced. Using Non Linear Autoregressive Neural Networks (NARX) with Matlab and Thingspeak IoT data capture, the results are favourable for open loop configuration although closed loop can be further improved, recommendations for the same are made at the end.