Domestic Hot Water Heating Prediction with the Utilization of Artificial Neural Network

Nowadays, there is a lot of pressure to adopt environmentally friendly energy sources like natural renewable energy. However, its instability and dependency on a variety of factors make their application a challenge. With the deployment of new technological systems, this unpredictability can be eliminated and their utilization maximized. This article discusses the possibility of using an artificially network to predict heat generation in a photovoltaic power plant. A Matlab script that creates a simple neural network serves as the foundation for the prediction model. The function fitting neural network is created using a backpropagation technique based on the Levenberg-Marquardt optimization. Because of the rapid response of the technique, it is suitable for basic predictions using fast computation speed. A simple artificial neural network based on the aforementioned approach is built in a publication. Created network is trained, tested, and validated on input and measured data. The constructed network is utilized as a prediction of heat generation after the results have been validated. The predictions' outcomes assist us to optimize consumption. This feature allows decreasing energy consumption from an external source. The utilization of a local source and optimization based on neural networks increase building energy efficiency The second important benefit is a decrease in the environmental impact and total carbon footprint of buildings.