Predictive Control of Hydronic Floor Heating Systems using Neural Networks and Genetic Algorithms

This paper presents the use a neural network and a micro genetic algorithm to
optimize future set-points in existing hydronic floor heating systems for improved energy efficiency. The neural network can be trained to predict the impact of changes in set-points on future room temperatures. Additionally, weather disturbances such as solar heat gain can be anticipated and compensated for, while taking into account the slow dynamics of the floor. Together with a genetic algorithm, they provide a way to search for optimal future set-point sequences, when convexity and continuity in the solution space is not guaranteed. Evaluation of the performance of multiple neural networks is performed, using different levels of information, and optimization results are presented on a detailed house simulation model.