An approximate integral model with an artificial neural network for heat exchangers
暂无分享,去创建一个
In order to improve predicting precision and increase computation speed of simulation for heat exchangers, a novel method is presented in this paper, whereby an approximate integral model is used to simplify the original distributed parameter model and an artificial neural network is combined to reflect the nonlinear relations. This model is applied in actual calculations of fin-and-tube condensers and high precision is achieved. Where the calculated outlet temperature of refrigerant and that of air, the average errors are both less than 0.2 °C. For the heat exchange of the condenser, the average error is less than 1 0.2 °C. For the heat exchange of the condenser, the average error is less than 1%. The calculation speed of the approximate integral model is two orders of magnitude faster than that of the distributed-parameter model. © 2004 Wiley Periodicals, Inc. Heat Trans Asian Res, 33(3): 153–160, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/htj.20006
[1] Guoliang Ding,et al. Steady-state simulation of screw liquid chillers , 2002 .
[2] Hao Liu,et al. A fast simulation model combining with artificial neural networks for fin‐and‐tube condenser , 2002 .
[3] Mihir Sen,et al. Simulation of heat exchanger performance by artificial neural networks , 1999 .
[4] S. Touber,et al. Distributed and non-steady-state modelling of an air cooler , 1991 .