دانلود Experimental Machine Learning Study on CO2 Gas Dispersion

Abstract

Machine    learning    (ML)    is    expending    its    application in many practical areas such as image recognition, natural language processing, games, etc. Simulated modeling of   gas   diffusion   can   be   one   of   the   applications.   This   experimental research was designed to know the potential of machine  learning  methods  in  modeling  CO2  gas  dispersion.  Dispersion data of gases can be collected with sensing devices so  that  ML-based  techniques  can  be  applied  to  simulate  the  diffusion.  In  this  study,  three  methods  were  explored  and  compared;   linear   interpolation,   Multi-Layer   Perceptron   (MLP)  and  Deep  Multi-Layer  Perceptron  (DLP).  A  set  of  experiments was conducted to collect dispersion data of CO2 gas. The experiments were executed in a wide room with two doors and eight windows that are enough to refresh the room air. Three sets of data were collected for learning and one set for  testing.  The  Root  Mean  Square  Deviation  (RMSD)  was  applied  to  compare  the  three  methods.  The  DLP  method  showed the lowest RMSD comparing with real test data, the linear interpolation the next and the MLP the last.