Partager

Estimation de la température du point de rosée de l'eau à l'aide de mécanismes d'intelligence artificielle.

Estimation of air dew point temperature using computational intelligence schemes.

Auteurs : BAGHBAN A., BAHADORI M., ROZYN J., et al.

Type d'article : Article

Résumé

The condensation of water vapor is a crucial problem, which might have serious problems, i.e. corrosion of metals and the wash out of protective coating of apparatuses, devices and pneumatic systems. Therefore, the dew point temperature of air at atmospheric pressure should be estimated with the intention of designing and applying the suitable kind of dryer. In the current contribution, two models based on statistical learning theories, i.e. Least Square Support Vector Machine (LSSVM) and Adaptive Neuro Fuzzy Inference System (ANFIS), were developed to predict the dew point temperature of moist air at atmospheric pressure over extensive range of temperature and relative humidity. Moreover, to optimize the corresponding parameters of these models, a Genetic Algorithm (GA) was applied. In this regard, a set of accessible data containing 1300 data points of moist air dew point in the temperature range of 0–50?°C, at a relative humidity up to 100%, and atmospheric pressure has been gathered from the reference. Estimations are found to be in excellent agreement with the reported data. The obtained values of Mean Squared Error (MSE) and R-Square (R2) were 0.000016, 1.0000 and 0.382402, 0.9987 for the LSSVM and ANFIS models respectively. The present tools can be of massive practical value for engineers and researchers as a quick check of the dew points of moist air.

Détails

  • Titre original : Estimation of air dew point temperature using computational intelligence schemes.
  • Identifiant de la fiche : 30017070
  • Langues : Anglais
  • Source : Applied Thermal Engineering - vol. 93
  • Date d'édition : 25/01/2016
  • DOI : http://dx.doi.org/10.1016/j.applthermaleng.2015.10.056

Liens


Voir d'autres articles du même numéro (72)
Voir la source

Indexation