Heterogeneous ensemble neural network for forecasting the state of multi-zone heating facilities

The research is aimed at increasing the accuracy of forecasting the state of multi-zone thermal facilities. Such facilities include multi-room premises, multi-zone greenhouses, tunnel kilns for brick production, and others. The high inertia of such facilities reduces the effectiveness of "ad hoc control". Modern proactive control systems based on forecasting are mainly based on using neural network training. However, to forecast the state of a specific multi-zone thermal facility, training the network requires a very large dataset, which is difficult to create and use. A combined neuro-structural method for forecasting the state of multi-zone thermal facilities is proposed, in which the structure of the neural model reflects the structure of the mutual influence of the facility zones. The research of the method has shown the possibility of ensuring sufficiently high forecast accuracy with a smaller size of the training dataset.