The application of artificial neural network for calculation of heat power consumption, using infrared system

摘要

1. The application of the infrared system and the neural networks in the research of the heat transfer process occurring between the radiator and the heated room, creates new possibilities of the analysis and control of this process. 2. The calculation of the heat power consumption in the steady state is possible using the described algorithm. 3. Use of neural network models, trained with the data obtained from experimental research gives the limiting error less or equal to 2%, if the measured heat power consumption lies within the range of the standard characteristic - Fig. 8. 4. The best simulation results were obtained for the network with four neurons in the hidden layer, trained with the Levenberg-Marquardt optimization algorithm.